sqlite-utils/sqlite_utils/db.py

from .utils import sqlite3, OperationalError, suggest_column_types, column_affinity
from collections import namedtuple, OrderedDict
from collections.abc import Mapping
import contextlib
import datetime
import decimal
import hashlib
import inspect
import itertools
import json
import os
import pathlib
import re
from sqlite_fts4 import rank_bm25
import sys
import textwrap
import uuid

SQLITE_MAX_VARS = 999

_virtual_table_using_re = re.compile(
    r"""
^ # Start of string
\s*CREATE\s+VIRTUAL\s+TABLE\s+ # CREATE VIRTUAL TABLE
(
    '(?P<squoted_table>[^']*(?:''[^']*)*)' | # single quoted name
    "(?P<dquoted_table>[^"]*(?:""[^"]*)*)" | # double quoted name
    `(?P<backtick_table>[^`]+)`            | # `backtick` quoted name
    \[(?P<squarequoted_table>[^\]]+)\]     | # [...] quoted name
    (?P<identifier>                          # SQLite non-quoted identifier
        [A-Za-z_\u0080-\uffff]  # \u0080-\uffff = "any character larger than u007f"
        [A-Za-z_\u0080-\uffff0-9\$]* # zero-or-more alphanemuric or $
    )
)
\s+(IF\s+NOT\s+EXISTS\s+)?      # IF NOT EXISTS (optional)
USING\s+(?P<using>\w+)          # e.g. USING FTS5
""",
    re.VERBOSE | re.IGNORECASE,
)

try:
    import pandas as pd
except ImportError:
    pd = None

try:
    import numpy as np
except ImportError:
    np = None

Column = namedtuple(
    "Column", ("cid", "name", "type", "notnull", "default_value", "is_pk")
)
ColumnDetails = namedtuple(
    "ColumnDetails",
    (
        "table",
        "column",
        "total_rows",
        "num_null",
        "num_blank",
        "num_distinct",
        "most_common",
        "least_common",
    ),
)
ForeignKey = namedtuple(
    "ForeignKey", ("table", "column", "other_table", "other_column")
)
Index = namedtuple("Index", ("seq", "name", "unique", "origin", "partial", "columns"))
Trigger = namedtuple("Trigger", ("name", "table", "sql"))


DEFAULT = object()

COLUMN_TYPE_MAPPING = {
    float: "FLOAT",
    int: "INTEGER",
    bool: "INTEGER",
    str: "TEXT",
    bytes.__class__: "BLOB",
    bytes: "BLOB",
    memoryview: "BLOB",
    datetime.datetime: "TEXT",
    datetime.date: "TEXT",
    datetime.time: "TEXT",
    decimal.Decimal: "FLOAT",
    None.__class__: "TEXT",
    uuid.UUID: "TEXT",
    # SQLite explicit types
    "TEXT": "TEXT",
    "INTEGER": "INTEGER",
    "FLOAT": "FLOAT",
    "BLOB": "BLOB",
    "text": "TEXT",
    "integer": "INTEGER",
    "float": "FLOAT",
    "blob": "BLOB",
}
# If numpy is available, add more types
if np:
    COLUMN_TYPE_MAPPING.update(
        {
            np.int8: "INTEGER",
            np.int16: "INTEGER",
            np.int32: "INTEGER",
            np.int64: "INTEGER",
            np.uint8: "INTEGER",
            np.uint16: "INTEGER",
            np.uint32: "INTEGER",
            np.uint64: "INTEGER",
            np.float16: "FLOAT",
            np.float32: "FLOAT",
            np.float64: "FLOAT",
        }
    )

# If pandas is available, add more types
if pd:
    COLUMN_TYPE_MAPPING.update({pd.Timestamp: "TEXT"})


class AlterError(Exception):
    pass


class NoObviousTable(Exception):
    pass


class BadPrimaryKey(Exception):
    pass


class NotFoundError(Exception):
    pass


class PrimaryKeyRequired(Exception):
    pass


class InvalidColumns(Exception):
    pass


_COUNTS_TABLE_CREATE_SQL = """
CREATE TABLE IF NOT EXISTS [{}](
   [table] TEXT PRIMARY KEY,
   count INTEGER DEFAULT 0
);
""".strip()


class Database:
    _counts_table_name = "_counts"
    use_counts_table = False

    def __init__(
        self,
        filename_or_conn=None,
        memory=False,
        recreate=False,
        recursive_triggers=True,
        tracer=None,
        use_counts_table=False,
    ):
        assert (filename_or_conn is not None and not memory) or (
            filename_or_conn is None and memory
        ), "Either specify a filename_or_conn or pass memory=True"
        if memory or filename_or_conn == ":memory:":
            self.conn = sqlite3.connect(":memory:")
        elif isinstance(filename_or_conn, (str, pathlib.Path)):
            if recreate and os.path.exists(filename_or_conn):
                os.remove(filename_or_conn)
            self.conn = sqlite3.connect(str(filename_or_conn))
        else:
            assert not recreate, "recreate cannot be used with connections, only paths"
            self.conn = filename_or_conn
        self._tracer = tracer
        if recursive_triggers:
            self.execute("PRAGMA recursive_triggers=on;")
        self._registered_functions = set()
        self.use_counts_table = use_counts_table

    @contextlib.contextmanager
    def tracer(self, tracer=None):
        prev_tracer = self._tracer
        self._tracer = tracer or print
        try:
            yield self
        finally:
            self._tracer = prev_tracer

    def __getitem__(self, table_name):
        return self.table(table_name)

    def __repr__(self):
        return "<Database {}>".format(self.conn)

    def register_function(self, fn=None, deterministic=None, replace=False):
        def register(fn):
            name = fn.__name__
            arity = len(inspect.signature(fn).parameters)
            if not replace and (name, arity) in self._registered_functions:
                return fn
            kwargs = {}
            if deterministic and sys.version_info >= (3, 8):
                kwargs["deterministic"] = True
            self.conn.create_function(name, arity, fn, **kwargs)
            self._registered_functions.add((name, arity))
            return fn

        if fn is None:
            return register
        else:
            register(fn)

    def register_fts4_bm25(self):
        self.register_function(rank_bm25, deterministic=True)

    def execute(self, sql, parameters=None):
        if self._tracer:
            self._tracer(sql, parameters)
        if parameters is not None:
            return self.conn.execute(sql, parameters)
        else:
            return self.conn.execute(sql)

    def executescript(self, sql):
        if self._tracer:
            self._tracer(sql, None)
        return self.conn.executescript(sql)

    def table(self, table_name, **kwargs):
        klass = View if table_name in self.view_names() else Table
        return klass(self, table_name, **kwargs)

    def quote(self, value):
        # Normally we would use .execute(sql, [params]) for escaping, but
        # occasionally that isn't available - most notable when we need
        # to include a "... DEFAULT 'value'" in a column definition.
        return self.execute(
            # Use SQLite itself to correctly escape this string:
            "SELECT quote(:value)",
            {"value": value},
        ).fetchone()[0]

    def table_names(self, fts4=False, fts5=False):
        where = ["type = 'table'"]
        if fts4:
            where.append("sql like '%USING FTS4%'")
        if fts5:
            where.append("sql like '%USING FTS5%'")
        sql = "select name from sqlite_master where {}".format(" AND ".join(where))
        return [r[0] for r in self.execute(sql).fetchall()]

    def view_names(self):
        return [
            r[0]
            for r in self.execute(
                "select name from sqlite_master where type = 'view'"
            ).fetchall()
        ]

    @property
    def tables(self):
        return [self[name] for name in self.table_names()]

    @property
    def views(self):
        return [self[name] for name in self.view_names()]

    @property
    def triggers(self):
        return [
            Trigger(*r)
            for r in self.execute(
                "select name, tbl_name, sql from sqlite_master where type = 'trigger'"
            ).fetchall()
        ]

    @property
    def triggers_dict(self):
        "Returns {trigger_name: sql} dictionary"
        return {trigger.name: trigger.sql for trigger in self.triggers}

    @property
    def journal_mode(self):
        return self.execute("PRAGMA journal_mode;").fetchone()[0]

    def enable_wal(self):
        if self.journal_mode != "wal":
            self.execute("PRAGMA journal_mode=wal;")

    def disable_wal(self):
        if self.journal_mode != "delete":
            self.execute("PRAGMA journal_mode=delete;")

    def _ensure_counts_table(self):
        with self.conn:
            self.execute(_COUNTS_TABLE_CREATE_SQL.format(self._counts_table_name))

    def enable_counts(self):
        self._ensure_counts_table()
        for table in self.tables:
            if (
                table.virtual_table_using is None
                and table.name != self._counts_table_name
            ):
                table.enable_counts()
        self.use_counts_table = True

    def cached_counts(self, tables=None):
        sql = "select [table], count from {}".format(self._counts_table_name)
        if tables:
            sql += " where [table] in ({})".format(", ".join("?" for table in tables))
        try:
            return {r[0]: r[1] for r in self.execute(sql, tables).fetchall()}
        except OperationalError:
            return {}

    def reset_counts(self):
        tables = [table for table in self.tables if table.has_counts_triggers]
        with self.conn:
            self._ensure_counts_table()
            counts_table = self[self._counts_table_name]
            counts_table.delete_where()
            counts_table.insert_all(
                {"table": table.name, "count": table.execute_count()}
                for table in tables
            )

    def execute_returning_dicts(self, sql, params=None):
        cursor = self.execute(sql, params or tuple())
        keys = [d[0] for d in cursor.description]
        return [dict(zip(keys, row)) for row in cursor.fetchall()]

    def resolve_foreign_keys(self, name, foreign_keys):
        # foreign_keys may be a list of strcolumn names, a list of ForeignKey tuples,
        # a list of tuple-pairs or a list of tuple-triples. We want to turn
        # it into a list of ForeignKey tuples
        if all(isinstance(fk, ForeignKey) for fk in foreign_keys):
            return foreign_keys
        if all(isinstance(fk, str) for fk in foreign_keys):
            # It's a list of columns
            fks = []
            for column in foreign_keys:
                other_table = self[name].guess_foreign_table(column)
                other_column = self[name].guess_foreign_column(other_table)
                fks.append(ForeignKey(name, column, other_table, other_column))
            return fks
        assert all(
            isinstance(fk, (tuple, list)) for fk in foreign_keys
        ), "foreign_keys= should be a list of tuples"
        fks = []
        for tuple_or_list in foreign_keys:
            assert len(tuple_or_list) in (
                2,
                3,
            ), "foreign_keys= should be a list of tuple pairs or triples"
            if len(tuple_or_list) == 3:
                fks.append(
                    ForeignKey(
                        name, tuple_or_list[0], tuple_or_list[1], tuple_or_list[2]
                    )
                )
            else:
                # Guess the primary key
                fks.append(
                    ForeignKey(
                        name,
                        tuple_or_list[0],
                        tuple_or_list[1],
                        self[name].guess_foreign_column(tuple_or_list[1]),
                    )
                )
        return fks

    def create_table_sql(
        self,
        name,
        columns,
        pk=None,
        foreign_keys=None,
        column_order=None,
        not_null=None,
        defaults=None,
        hash_id=None,
        extracts=None,
    ):
        foreign_keys = self.resolve_foreign_keys(name, foreign_keys or [])
        foreign_keys_by_column = {fk.column: fk for fk in foreign_keys}
        # any extracts will be treated as integer columns with a foreign key
        extracts = resolve_extracts(extracts)
        for extract_column, extract_table in extracts.items():
            if isinstance(extract_column, tuple):
                assert False
            # Ensure other table exists
            if not self[extract_table].exists():
                self.create_table(extract_table, {"id": int, "value": str}, pk="id")
            columns[extract_column] = int
            foreign_keys_by_column[extract_column] = ForeignKey(
                name, extract_column, extract_table, "id"
            )
        # Soundness check not_null, and defaults if provided
        not_null = not_null or set()
        defaults = defaults or {}
        assert all(
            n in columns for n in not_null
        ), "not_null set {} includes items not in columns {}".format(
            repr(not_null), repr(set(columns.keys()))
        )
        assert all(
            n in columns for n in defaults
        ), "defaults set {} includes items not in columns {}".format(
            repr(set(defaults)), repr(set(columns.keys()))
        )
        validate_column_names(columns.keys())
        column_items = list(columns.items())
        if column_order is not None:
            column_items.sort(
                key=lambda p: column_order.index(p[0]) if p[0] in column_order else 999
            )
        if hash_id:
            column_items.insert(0, (hash_id, str))
            pk = hash_id
        # Soundness check foreign_keys point to existing tables
        for fk in foreign_keys:
            if not any(
                c for c in self[fk.other_table].columns if c.name == fk.other_column
            ):
                raise AlterError(
                    "No such column: {}.{}".format(fk.other_table, fk.other_column)
                )

        column_defs = []
        # ensure pk is a tuple
        single_pk = None
        if isinstance(pk, list) and len(pk) == 1 and isinstance(pk[0], str):
            pk = pk[0]
        if isinstance(pk, str):
            single_pk = pk
            if pk not in [c[0] for c in column_items]:
                column_items.insert(0, (pk, int))
        for column_name, column_type in column_items:
            column_extras = []
            if column_name == single_pk:
                column_extras.append("PRIMARY KEY")
            if column_name in not_null:
                column_extras.append("NOT NULL")
            if column_name in defaults and defaults[column_name] is not None:
                column_extras.append(
                    "DEFAULT {}".format(self.quote(defaults[column_name]))
                )
            if column_name in foreign_keys_by_column:
                column_extras.append(
                    "REFERENCES [{other_table}]([{other_column}])".format(
                        other_table=foreign_keys_by_column[column_name].other_table,
                        other_column=foreign_keys_by_column[column_name].other_column,
                    )
                )
            column_defs.append(
                "   [{column_name}] {column_type}{column_extras}".format(
                    column_name=column_name,
                    column_type=COLUMN_TYPE_MAPPING[column_type],
                    column_extras=(" " + " ".join(column_extras))
                    if column_extras
                    else "",
                )
            )
        extra_pk = ""
        if single_pk is None and pk and len(pk) > 1:
            extra_pk = ",\n   PRIMARY KEY ({pks})".format(
                pks=", ".join(["[{}]".format(p) for p in pk])
            )
        columns_sql = ",\n".join(column_defs)
        sql = """CREATE TABLE [{table}] (
{columns_sql}{extra_pk}
);
        """.format(
            table=name, columns_sql=columns_sql, extra_pk=extra_pk
        )
        return sql

    def create_table(
        self,
        name,
        columns,
        pk=None,
        foreign_keys=None,
        column_order=None,
        not_null=None,
        defaults=None,
        hash_id=None,
        extracts=None,
    ):
        sql = self.create_table_sql(
            name=name,
            columns=columns,
            pk=pk,
            foreign_keys=foreign_keys,
            column_order=column_order,
            not_null=not_null,
            defaults=defaults,
            hash_id=hash_id,
            extracts=extracts,
        )
        self.execute(sql)
        return self.table(
            name,
            pk=pk,
            foreign_keys=foreign_keys,
            column_order=column_order,
            not_null=not_null,
            defaults=defaults,
            hash_id=hash_id,
        )

    def create_view(self, name, sql, ignore=False, replace=False):
        assert not (
            ignore and replace
        ), "Use one or the other of ignore/replace, not both"
        create_sql = "CREATE VIEW {name} AS {sql}".format(name=name, sql=sql)
        if ignore or replace:
            # Does view exist already?
            if name in self.view_names():
                if ignore:
                    return self
                elif replace:
                    # If SQL is the same, do nothing
                    if create_sql == self[name].schema:
                        return self
                    self[name].drop()
        self.execute(create_sql)
        return self

    def m2m_table_candidates(self, table, other_table):
        "Returns potential m2m tables for arguments, based on FKs"
        candidates = []
        tables = {table, other_table}
        for table in self.tables:
            # Does it have foreign keys to both table and other_table?
            has_fks_to = {fk.other_table for fk in table.foreign_keys}
            if has_fks_to.issuperset(tables):
                candidates.append(table.name)
        return candidates

    def add_foreign_keys(self, foreign_keys):
        # foreign_keys is a list of explicit 4-tuples
        assert all(
            len(fk) == 4 and isinstance(fk, (list, tuple)) for fk in foreign_keys
        ), "foreign_keys must be a list of 4-tuples, (table, column, other_table, other_column)"

        foreign_keys_to_create = []

        # Verify that all tables and columns exist
        for table, column, other_table, other_column in foreign_keys:
            if not self[table].exists():
                raise AlterError("No such table: {}".format(table))
            if column not in self[table].columns_dict:
                raise AlterError("No such column: {} in {}".format(column, table))
            if not self[other_table].exists():
                raise AlterError("No such other_table: {}".format(other_table))
            if (
                other_column != "rowid"
                and other_column not in self[other_table].columns_dict
            ):
                raise AlterError(
                    "No such other_column: {} in {}".format(other_column, other_table)
                )
            # We will silently skip foreign keys that exist already
            if not any(
                fk
                for fk in self[table].foreign_keys
                if fk.column == column
                and fk.other_table == other_table
                and fk.other_column == other_column
            ):
                foreign_keys_to_create.append(
                    (table, column, other_table, other_column)
                )

        # Construct SQL for use with "UPDATE sqlite_master SET sql = ? WHERE name = ?"
        table_sql = {}
        for table, column, other_table, other_column in foreign_keys_to_create:
            old_sql = table_sql.get(table, self[table].schema)
            extra_sql = ",\n   FOREIGN KEY({column}) REFERENCES {other_table}({other_column})\n".format(
                column=column, other_table=other_table, other_column=other_column
            )
            # Stick that bit in at the very end just before the closing ')'
            last_paren = old_sql.rindex(")")
            new_sql = old_sql[:last_paren].strip() + extra_sql + old_sql[last_paren:]
            table_sql[table] = new_sql

        # And execute it all within a single transaction
        with self.conn:
            cursor = self.conn.cursor()
            schema_version = cursor.execute("PRAGMA schema_version").fetchone()[0]
            cursor.execute("PRAGMA writable_schema = 1")
            for table_name, new_sql in table_sql.items():
                cursor.execute(
                    "UPDATE sqlite_master SET sql = ? WHERE name = ?",
                    (new_sql, table_name),
                )
            cursor.execute("PRAGMA schema_version = %d" % (schema_version + 1))
            cursor.execute("PRAGMA writable_schema = 0")
        # Have to VACUUM outside the transaction to ensure .foreign_keys property
        # can see the newly created foreign key.
        self.vacuum()

    def index_foreign_keys(self):
        for table_name in self.table_names():
            table = self[table_name]
            existing_indexes = {
                i.columns[0] for i in table.indexes if len(i.columns) == 1
            }
            for fk in table.foreign_keys:
                if fk.column not in existing_indexes:
                    table.create_index([fk.column])

    def vacuum(self):
        self.execute("VACUUM;")


class Queryable:
    def exists(self):
        return False

    def __init__(self, db, name):
        self.db = db
        self.name = name

    def execute_count(self):
        return self.db.execute(
            "select count(*) from [{}]".format(self.name)
        ).fetchone()[0]

    @property
    def count(self):
        return self.execute_count()

    @property
    def rows(self):
        return self.rows_where()

    def rows_where(
        self,
        where=None,
        where_args=None,
        order_by=None,
        select="*",
        limit=None,
        offset=None,
    ):
        if not self.exists():
            return []
        sql = "select {} from [{}]".format(select, self.name)
        if where is not None:
            sql += " where " + where
        if order_by is not None:
            sql += " order by " + order_by
        if limit is not None:
            sql += " limit {}".format(limit)
        if offset is not None:
            sql += " offset {}".format(offset)
        cursor = self.db.execute(sql, where_args or [])
        columns = [c[0] for c in cursor.description]
        for row in cursor:
            yield dict(zip(columns, row))

    @property
    def columns(self):
        if not self.exists():
            return []
        rows = self.db.execute("PRAGMA table_info([{}])".format(self.name)).fetchall()
        return [Column(*row) for row in rows]

    @property
    def columns_dict(self):
        "Returns {column: python-type} dictionary"
        return {column.name: column_affinity(column.type) for column in self.columns}

    @property
    def schema(self):
        return self.db.execute(
            "select sql from sqlite_master where name = ?", (self.name,)
        ).fetchone()[0]


class Table(Queryable):
    last_rowid = None
    last_pk = None

    def __init__(
        self,
        db,
        name,
        pk=None,
        foreign_keys=None,
        column_order=None,
        not_null=None,
        defaults=None,
        batch_size=100,
        hash_id=None,
        alter=False,
        ignore=False,
        replace=False,
        extracts=None,
        conversions=None,
        columns=None,
    ):
        super().__init__(db, name)
        self._defaults = dict(
            pk=pk,
            foreign_keys=foreign_keys,
            column_order=column_order,
            not_null=not_null,
            defaults=defaults,
            batch_size=batch_size,
            hash_id=hash_id,
            alter=alter,
            ignore=ignore,
            replace=replace,
            extracts=extracts,
            conversions=conversions or {},
            columns=columns,
        )

    def __repr__(self):
        return "<Table {}{}>".format(
            self.name,
            " (does not exist yet)"
            if not self.exists()
            else " ({})".format(", ".join(c.name for c in self.columns)),
        )

    @property
    def count(self):
        if self.db.use_counts_table:
            counts = self.db.cached_counts([self.name])
            if counts:
                return next(iter(counts.values()))
        return self.execute_count()

    def exists(self):
        return self.name in self.db.table_names()

    @property
    def pks(self):
        names = [column.name for column in self.columns if column.is_pk]
        if not names:
            names = ["rowid"]
        return names

    def get(self, pk_values):
        if not isinstance(pk_values, (list, tuple)):
            pk_values = [pk_values]
        pks = self.pks
        last_pk = pk_values[0] if len(pks) == 1 else pk_values
        if len(pks) != len(pk_values):
            raise NotFoundError(
                "Need {} primary key value{}".format(
                    len(pks), "" if len(pks) == 1 else "s"
                )
            )

        wheres = ["[{}] = ?".format(pk_name) for pk_name in pks]
        rows = self.rows_where(" and ".join(wheres), pk_values)
        try:
            row = list(rows)[0]
            self.last_pk = last_pk
            return row
        except IndexError:
            raise NotFoundError

    @property
    def foreign_keys(self):
        fks = []
        for row in self.db.execute(
            "PRAGMA foreign_key_list([{}])".format(self.name)
        ).fetchall():
            if row is not None:
                id, seq, table_name, from_, to_, on_update, on_delete, match = row
                fks.append(
                    ForeignKey(
                        table=self.name,
                        column=from_,
                        other_table=table_name,
                        other_column=to_,
                    )
                )
        return fks

    @property
    def virtual_table_using(self):
        "Returns type of virtual table or None if this is not a virtual table"
        match = _virtual_table_using_re.match(self.schema)
        if match is None:
            return None
        return match.groupdict()["using"].upper()

    @property
    def indexes(self):
        sql = 'PRAGMA index_list("{}")'.format(self.name)
        indexes = []
        for row in self.db.execute_returning_dicts(sql):
            index_name = row["name"]
            index_name_quoted = (
                '"{}"'.format(index_name)
                if not index_name.startswith('"')
                else index_name
            )
            column_sql = "PRAGMA index_info({})".format(index_name_quoted)
            columns = []
            for seqno, cid, name in self.db.execute(column_sql).fetchall():
                columns.append(name)
            row["columns"] = columns
            # These columns may be missing on older SQLite versions:
            for key, default in {"origin": "c", "partial": 0}.items():
                if key not in row:
                    row[key] = default
            indexes.append(Index(**row))
        return indexes

    @property
    def triggers(self):
        return [
            Trigger(*r)
            for r in self.db.execute(
                "select name, tbl_name, sql from sqlite_master where type = 'trigger'"
                " and tbl_name = ?",
                (self.name,),
            ).fetchall()
        ]

    @property
    def triggers_dict(self):
        "Returns {trigger_name: sql} dictionary"
        return {trigger.name: trigger.sql for trigger in self.triggers}

    def create(
        self,
        columns,
        pk=None,
        foreign_keys=None,
        column_order=None,
        not_null=None,
        defaults=None,
        hash_id=None,
        extracts=None,
    ):
        columns = {name: value for (name, value) in columns.items()}
        with self.db.conn:
            self.db.create_table(
                self.name,
                columns,
                pk=pk,
                foreign_keys=foreign_keys,
                column_order=column_order,
                not_null=not_null,
                defaults=defaults,
                hash_id=hash_id,
                extracts=extracts,
            )
        return self

    def transform(
        self,
        *,
        types=None,
        rename=None,
        drop=None,
        pk=DEFAULT,
        not_null=None,
        defaults=None,
        drop_foreign_keys=None,
        column_order=None,
    ):
        assert self.exists(), "Cannot transform a table that doesn't exist yet"
        sqls = self.transform_sql(
            types=types,
            rename=rename,
            drop=drop,
            pk=pk,
            not_null=not_null,
            defaults=defaults,
            drop_foreign_keys=drop_foreign_keys,
            column_order=column_order,
        )
        pragma_foreign_keys_was_on = self.db.execute("PRAGMA foreign_keys").fetchone()[
            0
        ]
        try:
            if pragma_foreign_keys_was_on:
                self.db.execute("PRAGMA foreign_keys=0;")
            with self.db.conn:
                for sql in sqls:
                    self.db.execute(sql)
                # Run the foreign_key_check before we commit
                if pragma_foreign_keys_was_on:
                    self.db.execute("PRAGMA foreign_key_check;")
        finally:
            if pragma_foreign_keys_was_on:
                self.db.execute("PRAGMA foreign_keys=1;")
        return self

    def transform_sql(
        self,
        *,
        types=None,
        rename=None,
        drop=None,
        pk=DEFAULT,
        not_null=None,
        defaults=None,
        drop_foreign_keys=None,
        column_order=None,
        tmp_suffix=None,
    ):
        types = types or {}
        rename = rename or {}
        drop = drop or set()
        new_table_name = "{}_new_{}".format(
            self.name, tmp_suffix or os.urandom(6).hex()
        )
        current_column_pairs = list(self.columns_dict.items())
        new_column_pairs = []
        copy_from_to = {column: column for column, _ in current_column_pairs}
        for name, type_ in current_column_pairs:
            type_ = types.get(name) or type_
            if name in drop:
                del [copy_from_to[name]]
                continue
            new_name = rename.get(name) or name
            new_column_pairs.append((new_name, type_))
            copy_from_to[name] = new_name

        sqls = []
        if pk is DEFAULT:
            pks_renamed = tuple(rename.get(p) or p for p in self.pks)
            if len(pks_renamed) == 1:
                pk = pks_renamed[0]
            else:
                pk = pks_renamed

        # not_null may be a set or dict, need to convert to a set
        create_table_not_null = {
            rename.get(c.name) or c.name
            for c in self.columns
            if c.notnull
            if c.name not in drop
        }
        if isinstance(not_null, dict):
            # Remove any columns with a value of False
            for key, value in not_null.items():
                # Column may have been renamed
                key = rename.get(key) or key
                if value is False and key in create_table_not_null:
                    create_table_not_null.remove(key)
                else:
                    create_table_not_null.add(key)
        elif isinstance(not_null, set):
            create_table_not_null.update((rename.get(k) or k) for k in not_null)
        elif not_null is None:
            pass
        else:
            assert False, "not_null must be a dict or a set or None"
        # defaults=
        create_table_defaults = {
            (rename.get(c.name) or c.name): c.default_value
            for c in self.columns
            if c.default_value is not None and c.name not in drop
        }
        if defaults is not None:
            create_table_defaults.update(
                {rename.get(c) or c: v for c, v in defaults.items()}
            )

        # foreign_keys
        create_table_foreign_keys = []
        for table, column, other_table, other_column in self.foreign_keys:
            if (drop_foreign_keys is None) or (column not in drop_foreign_keys):
                create_table_foreign_keys.append(
                    (rename.get(column) or column, other_table, other_column)
                )

        if column_order is not None:
            column_order = [rename.get(col) or col for col in column_order]

        sqls.append(
            self.db.create_table_sql(
                new_table_name,
                dict(new_column_pairs),
                pk=pk,
                not_null=create_table_not_null,
                defaults=create_table_defaults,
                foreign_keys=create_table_foreign_keys,
                column_order=column_order,
            ).strip()
        )

        # Copy across data, respecting any renamed columns
        new_cols = []
        old_cols = []
        for from_, to_ in copy_from_to.items():
            old_cols.append(from_)
            new_cols.append(to_)
        copy_sql = "INSERT INTO [{new_table}] ({new_cols})\n   SELECT {old_cols} FROM [{old_table}];".format(
            new_table=new_table_name,
            old_table=self.name,
            old_cols=", ".join("[{}]".format(col) for col in old_cols),
            new_cols=", ".join("[{}]".format(col) for col in new_cols),
        )
        sqls.append(copy_sql)
        # Drop the old table
        sqls.append("DROP TABLE [{}];".format(self.name))
        # Rename the new one
        sqls.append(
            "ALTER TABLE [{}] RENAME TO [{}];".format(new_table_name, self.name)
        )
        return sqls

    def extract(self, columns, table=None, fk_column=None, rename=None):
        rename = rename or {}
        if isinstance(columns, str):
            columns = [columns]
        if not set(columns).issubset(self.columns_dict.keys()):
            raise InvalidColumns(
                "Invalid columns {} for table with columns {}".format(
                    columns, list(self.columns_dict.keys())
                )
            )
        table = table or "_".join(columns)
        first_column = columns[0]
        pks = self.pks
        lookup_table = self.db[table]
        fk_column = fk_column or "{}_id".format(table)
        magic_lookup_column = "{}_{}".format(fk_column, os.urandom(6).hex())

        # Populate the lookup table with all of the extracted unique values
        lookup_columns_definition = {
            (rename.get(col) or col): typ
            for col, typ in self.columns_dict.items()
            if col in columns
        }
        if lookup_table.exists():
            if not set(lookup_columns_definition.items()).issubset(
                lookup_table.columns_dict.items()
            ):
                raise InvalidColumns(
                    "Lookup table {} already exists but does not have columns {}".format(
                        table, lookup_columns_definition
                    )
                )
        else:
            lookup_table.create(
                {
                    **{
                        "id": int,
                    },
                    **lookup_columns_definition,
                },
                pk="id",
            )
        lookup_columns = [(rename.get(col) or col) for col in columns]
        lookup_table.create_index(lookup_columns, unique=True, if_not_exists=True)
        self.db.execute(
            "INSERT OR IGNORE INTO [{lookup_table}] ({lookup_columns}) SELECT DISTINCT {table_cols} FROM [{table}]".format(
                lookup_table=table,
                lookup_columns=", ".join("[{}]".format(c) for c in lookup_columns),
                table_cols=", ".join("[{}]".format(c) for c in columns),
                table=self.name,
            )
        )

        # Now add the new fk_column
        self.add_column(magic_lookup_column, int)

        # And populate it
        self.db.execute(
            "UPDATE [{table}] SET [{magic_lookup_column}] = (SELECT id FROM [{lookup_table}] WHERE {where})".format(
                table=self.name,
                magic_lookup_column=magic_lookup_column,
                lookup_table=table,
                where=" AND ".join(
                    "[{table}].[{column}] = [{lookup_table}].[{lookup_column}]".format(
                        table=self.name,
                        lookup_table=table,
                        column=column,
                        lookup_column=rename.get(column) or column,
                    )
                    for column in columns
                ),
            )
        )
        # Figure out the right column order
        column_order = []
        for c in self.columns:
            if c.name in columns and magic_lookup_column not in column_order:
                column_order.append(magic_lookup_column)
            elif c.name == magic_lookup_column:
                continue
            else:
                column_order.append(c.name)

        # Drop the unnecessary columns and rename lookup column
        self.transform(
            drop=set(columns),
            rename={magic_lookup_column: fk_column},
            column_order=column_order,
        )

        # And add the foreign key constraint
        self.add_foreign_key(fk_column, table, "id")
        return self

    def create_index(self, columns, index_name=None, unique=False, if_not_exists=False):
        if index_name is None:
            index_name = "idx_{}_{}".format(
                self.name.replace(" ", "_"), "_".join(columns)
            )
        sql = (
            textwrap.dedent(
                """
            CREATE {unique}INDEX {if_not_exists}[{index_name}]
                ON [{table_name}] ({columns});
        """
            )
            .strip()
            .format(
                index_name=index_name,
                table_name=self.name,
                columns=", ".join("[{}]".format(c) for c in columns),
                unique="UNIQUE " if unique else "",
                if_not_exists="IF NOT EXISTS " if if_not_exists else "",
            )
        )
        self.db.execute(sql)
        return self

    def add_column(
        self, col_name, col_type=None, fk=None, fk_col=None, not_null_default=None
    ):
        fk_col_type = None
        if fk is not None:
            # fk must be a valid table
            if not fk in self.db.table_names():
                raise AlterError("table '{}' does not exist".format(fk))
            # if fk_col specified, must be a valid column
            if fk_col is not None:
                if fk_col not in self.db[fk].columns_dict:
                    raise AlterError("table '{}' has no column {}".format(fk, fk_col))
            else:
                # automatically set fk_col to first primary_key of fk table
                pks = [c for c in self.db[fk].columns if c.is_pk]
                if pks:
                    fk_col = pks[0].name
                    fk_col_type = pks[0].type
                else:
                    fk_col = "rowid"
                    fk_col_type = "INTEGER"
        if col_type is None:
            col_type = str
        not_null_sql = None
        if not_null_default is not None:
            not_null_sql = "NOT NULL DEFAULT {}".format(self.db.quote(not_null_default))
        sql = "ALTER TABLE [{table}] ADD COLUMN [{col_name}] {col_type}{not_null_default};".format(
            table=self.name,
            col_name=col_name,
            col_type=fk_col_type or COLUMN_TYPE_MAPPING[col_type],
            not_null_default=(" " + not_null_sql) if not_null_sql else "",
        )
        self.db.execute(sql)
        if fk is not None:
            self.add_foreign_key(col_name, fk, fk_col)
        return self

    def drop(self):
        self.db.execute("DROP TABLE [{}]".format(self.name))

    def guess_foreign_table(self, column):
        column = column.lower()
        possibilities = [column]
        if column.endswith("_id"):
            column_without_id = column[:-3]
            possibilities.append(column_without_id)
            if not column_without_id.endswith("s"):
                possibilities.append(column_without_id + "s")
        elif not column.endswith("s"):
            possibilities.append(column + "s")
        existing_tables = {t.lower(): t for t in self.db.table_names()}
        for table in possibilities:
            if table in existing_tables:
                return existing_tables[table]
        # If we get here there's no obvious candidate - raise an error
        raise NoObviousTable(
            "No obvious foreign key table for column '{}' - tried {}".format(
                column, repr(possibilities)
            )
        )

    def guess_foreign_column(self, other_table):
        pks = [c for c in self.db[other_table].columns if c.is_pk]
        if len(pks) != 1:
            raise BadPrimaryKey(
                "Could not detect single primary key for table '{}'".format(other_table)
            )
        else:
            return pks[0].name

    def add_foreign_key(
        self, column, other_table=None, other_column=None, ignore=False
    ):
        # Ensure column exists
        if column not in self.columns_dict:
            raise AlterError("No such column: {}".format(column))
        # If other_table is not specified, attempt to guess it from the column
        if other_table is None:
            other_table = self.guess_foreign_table(column)
        # If other_column is not specified, detect the primary key on other_table
        if other_column is None:
            other_column = self.guess_foreign_column(other_table)

        # Soundness check that the other column exists
        if (
            not [c for c in self.db[other_table].columns if c.name == other_column]
            and other_column != "rowid"
        ):
            raise AlterError("No such column: {}.{}".format(other_table, other_column))
        # Check we do not already have an existing foreign key
        if any(
            fk
            for fk in self.foreign_keys
            if fk.column == column
            and fk.other_table == other_table
            and fk.other_column == other_column
        ):
            if ignore:
                return self
            else:
                raise AlterError(
                    "Foreign key already exists for {} => {}.{}".format(
                        column, other_table, other_column
                    )
                )
        self.db.add_foreign_keys([(self.name, column, other_table, other_column)])
        return self

    def enable_counts(self):
        sql = (
            textwrap.dedent(
                """
        {create_counts_table}
        CREATE TRIGGER IF NOT EXISTS [{table}{counts_table}_insert] AFTER INSERT ON [{table}]
        BEGIN
            INSERT OR REPLACE INTO [{counts_table}]
            VALUES (
                {table_quoted},
                COALESCE(
                    (SELECT count FROM [{counts_table}] WHERE [table] = {table_quoted}),
                0
                ) + 1
            );
        END;
        CREATE TRIGGER IF NOT EXISTS [{table}{counts_table}_delete] AFTER DELETE ON [{table}]
        BEGIN
            INSERT OR REPLACE INTO [{counts_table}]
            VALUES (
                {table_quoted},
                COALESCE(
                    (SELECT count FROM [{counts_table}] WHERE [table] = {table_quoted}),
                0
                ) - 1
            );
        END;
        INSERT OR REPLACE INTO _counts VALUES ({table_quoted}, (select count(*) from [{table}]));
        """
            )
            .strip()
            .format(
                create_counts_table=_COUNTS_TABLE_CREATE_SQL.format(
                    self.db._counts_table_name
                ),
                counts_table=self.db._counts_table_name,
                table=self.name,
                table_quoted=self.db.quote(self.name),
            )
        )
        with self.db.conn:
            self.db.conn.executescript(sql)
        self.db.use_counts_table = True

    @property
    def has_counts_triggers(self):
        trigger_names = {
            "{table}{counts_table}_{suffix}".format(
                counts_table=self.db._counts_table_name, table=self.name, suffix=suffix
            )
            for suffix in ["insert", "delete"]
        }
        return trigger_names.issubset(self.triggers_dict.keys())

    def enable_fts(
        self,
        columns,
        fts_version="FTS5",
        create_triggers=False,
        tokenize=None,
        replace=False,
    ):
        "Enables FTS on the specified columns."
        create_fts_sql = (
            textwrap.dedent(
                """
            CREATE VIRTUAL TABLE [{table}_fts] USING {fts_version} (
                {columns},{tokenize}
                content=[{table}]
            )
        """
            )
            .strip()
            .format(
                table=self.name,
                columns=", ".join("[{}]".format(c) for c in columns),
                fts_version=fts_version,
                tokenize="\n    tokenize='{}',".format(tokenize) if tokenize else "",
            )
        )
        should_recreate = False
        if replace and self.db["{}_fts".format(self.name)].exists():
            # Does the table need to be recreated?
            fts_schema = self.db["{}_fts".format(self.name)].schema
            if fts_schema != create_fts_sql:
                should_recreate = True
            expected_triggers = {self.name + suffix for suffix in ("_ai", "_ad", "_au")}
            existing_triggers = {t.name for t in self.triggers}
            has_triggers = existing_triggers.issuperset(expected_triggers)
            if has_triggers != create_triggers:
                should_recreate = True
            if not should_recreate:
                # Table with correct configuration already exists
                return self

        if should_recreate:
            self.disable_fts()

        self.db.executescript(create_fts_sql)
        self.populate_fts(columns)

        if create_triggers:
            old_cols = ", ".join("old.[{}]".format(c) for c in columns)
            new_cols = ", ".join("new.[{}]".format(c) for c in columns)
            triggers = (
                textwrap.dedent(
                    """
                CREATE TRIGGER [{table}_ai] AFTER INSERT ON [{table}] BEGIN
                  INSERT INTO [{table}_fts] (rowid, {columns}) VALUES (new.rowid, {new_cols});
                END;
                CREATE TRIGGER [{table}_ad] AFTER DELETE ON [{table}] BEGIN
                  INSERT INTO [{table}_fts] ([{table}_fts], rowid, {columns}) VALUES('delete', old.rowid, {old_cols});
                END;
                CREATE TRIGGER [{table}_au] AFTER UPDATE ON [{table}] BEGIN
                  INSERT INTO [{table}_fts] ([{table}_fts], rowid, {columns}) VALUES('delete', old.rowid, {old_cols});
                  INSERT INTO [{table}_fts] (rowid, {columns}) VALUES (new.rowid, {new_cols});
                END;
            """
                )
                .strip()
                .format(
                    table=self.name,
                    columns=", ".join("[{}]".format(c) for c in columns),
                    old_cols=old_cols,
                    new_cols=new_cols,
                )
            )
            self.db.executescript(triggers)
        return self

    def populate_fts(self, columns):
        sql = (
            textwrap.dedent(
                """
            INSERT INTO [{table}_fts] (rowid, {columns})
                SELECT rowid, {columns} FROM [{table}];
        """
            )
            .strip()
            .format(
                table=self.name, columns=", ".join("[{}]".format(c) for c in columns)
            )
        )
        self.db.executescript(sql)
        return self

    def disable_fts(self):
        fts_table = self.detect_fts()
        if fts_table:
            self.db[fts_table].drop()
        # Now delete the triggers that related to that table
        sql = (
            textwrap.dedent(
                """
            SELECT name FROM sqlite_master
                WHERE type = 'trigger'
                AND sql LIKE '% INSERT INTO [{}]%'
        """
            )
            .strip()
            .format(fts_table)
        )
        trigger_names = []
        for row in self.db.execute(sql).fetchall():
            trigger_names.append(row[0])
        with self.db.conn:
            for trigger_name in trigger_names:
                self.db.execute("DROP TRIGGER IF EXISTS [{}]".format(trigger_name))
        return self

    def rebuild_fts(self):
        fts_table = self.detect_fts()
        if fts_table is None:
            # Assume this is itself an FTS table
            fts_table = self.name
        self.db.execute(
            "INSERT INTO [{table}]([{table}]) VALUES('rebuild');".format(
                table=fts_table
            )
        )
        return self

    def detect_fts(self):
        "Detect if table has a corresponding FTS virtual table and return it"
        sql = (
            textwrap.dedent(
                """
            SELECT name FROM sqlite_master
                WHERE rootpage = 0
                AND (
                    sql LIKE '%VIRTUAL TABLE%USING FTS%content=%{table}%'
                    OR (
                        tbl_name = "{table}"
                        AND sql LIKE '%VIRTUAL TABLE%USING FTS%'
                    )
                )
        """
            )
            .strip()
            .format(table=self.name)
        )
        rows = self.db.execute(sql).fetchall()
        if len(rows) == 0:
            return None
        else:
            return rows[0][0]

    def optimize(self):
        fts_table = self.detect_fts()
        if fts_table is not None:
            self.db.execute(
                """
                INSERT INTO [{table}] ([{table}]) VALUES ("optimize");
            """.strip().format(
                    table=fts_table
                )
            )
        return self

    def search_sql(self, columns=None, order_by=None, limit=None, offset=None):
        # Pick names for table and rank column that don't clash
        original = "original_" if self.name == "original" else "original"
        columns_sql = "*"
        columns_with_prefix_sql = "[{}].*".format(original)
        if columns:
            columns_sql = ",\n        ".join("[{}]".format(c) for c in columns)
            columns_with_prefix_sql = ",\n    ".join(
                "[{}].[{}]".format(original, c) for c in columns
            )
        fts_table = self.detect_fts()
        assert fts_table, "Full-text search is not configured for table '{}'".format(
            self.name
        )
        virtual_table_using = self.db[fts_table].virtual_table_using
        sql = textwrap.dedent(
            """
        with {original} as (
            select
                rowid,
                {columns}
            from [{dbtable}]
        )
        select
            {columns_with_prefix}
        from
            [{original}]
            join [{fts_table}] on [{original}].rowid = [{fts_table}].rowid
        where
            [{fts_table}] match :query
        order by
            {order_by}
        {limit_offset}
        """
        ).strip()
        if virtual_table_using == "FTS5":
            rank_implementation = "[{}].rank".format(fts_table)
        else:
            self.db.register_fts4_bm25()
            rank_implementation = "rank_bm25(matchinfo([{}], 'pcnalx'))".format(
                fts_table
            )
        limit_offset = ""
        if limit is not None:
            limit_offset += " limit {}".format(limit)
        if offset is not None:
            limit_offset += " offset {}".format(offset)
        return sql.format(
            dbtable=self.name,
            original=original,
            columns=columns_sql,
            columns_with_prefix=columns_with_prefix_sql,
            fts_table=fts_table,
            order_by=order_by or rank_implementation,
            limit_offset=limit_offset.strip(),
        ).strip()

    def search(self, q, order_by=None, columns=None, limit=None, offset=None):
        cursor = self.db.execute(
            self.search_sql(
                order_by=order_by,
                columns=columns,
                limit=limit,
                offset=offset,
            ),
            {"query": q},
        )
        columns = [c[0] for c in cursor.description]
        for row in cursor:
            yield dict(zip(columns, row))

    def value_or_default(self, key, value):
        return self._defaults[key] if value is DEFAULT else value

    def delete(self, pk_values):
        if not isinstance(pk_values, (list, tuple)):
            pk_values = [pk_values]
        self.get(pk_values)
        wheres = ["[{}] = ?".format(pk_name) for pk_name in self.pks]
        sql = "delete from [{table}] where {wheres}".format(
            table=self.name, wheres=" and ".join(wheres)
        )
        with self.db.conn:
            self.db.execute(sql, pk_values)
        return self

    def delete_where(self, where=None, where_args=None):
        if not self.exists():
            return []
        sql = "delete from [{}]".format(self.name)
        if where is not None:
            sql += " where " + where
        self.db.execute(sql, where_args or [])
        return self

    def update(self, pk_values, updates=None, alter=False, conversions=None):
        updates = updates or {}
        conversions = conversions or {}
        if not isinstance(pk_values, (list, tuple)):
            pk_values = [pk_values]
        # Soundness check that the record exists (raises error if not):
        self.get(pk_values)
        if not updates:
            return self
        args = []
        sets = []
        wheres = []
        pks = self.pks
        validate_column_names(updates.keys())
        for key, value in updates.items():
            sets.append("[{}] = {}".format(key, conversions.get(key, "?")))
            args.append(jsonify_if_needed(value))
        wheres = ["[{}] = ?".format(pk_name) for pk_name in pks]
        args.extend(pk_values)
        sql = "update [{table}] set {sets} where {wheres}".format(
            table=self.name, sets=", ".join(sets), wheres=" and ".join(wheres)
        )
        with self.db.conn:
            try:
                rowcount = self.db.execute(sql, args).rowcount
            except OperationalError as e:
                if alter and (" column" in e.args[0]):
                    # Attempt to add any missing columns, then try again
                    self.add_missing_columns([updates])
                    rowcount = self.db.execute(sql, args).rowcount
                else:
                    raise

            # TODO: Test this works (rolls back) - use better exception:
            assert rowcount == 1
        self.last_pk = pk_values[0] if len(pks) == 1 else pk_values
        return self

    def build_insert_queries_and_params(
        self,
        extracts,
        chunk,
        all_columns,
        hash_id,
        upsert,
        pk,
        conversions,
        num_records_processed,
        replace,
        ignore,
    ):
        # values is the list of insert data that is passed to the
        # .execute() method - but some of them may be replaced by
        # new primary keys if we are extracting any columns.
        values = []
        extracts = resolve_extracts(extracts)
        for record in chunk:
            record_values = []
            for key in all_columns:
                value = jsonify_if_needed(
                    record.get(key, None if key != hash_id else _hash(record))
                )
                if key in extracts:
                    extract_table = extracts[key]
                    value = self.db[extract_table].lookup({"value": value})
                record_values.append(value)
            values.append(record_values)

        queries_and_params = []
        if upsert:
            if isinstance(pk, str):
                pks = [pk]
            else:
                pks = pk
            self.last_pk = None
            for record_values in values:
                # TODO: make more efficient:
                record = dict(zip(all_columns, record_values))
                sql = "INSERT OR IGNORE INTO [{table}]({pks}) VALUES({pk_placeholders});".format(
                    table=self.name,
                    pks=", ".join(["[{}]".format(p) for p in pks]),
                    pk_placeholders=", ".join(["?" for p in pks]),
                )
                queries_and_params.append((sql, [record[col] for col in pks]))
                # UPDATE [book] SET [name] = 'Programming' WHERE [id] = 1001;
                set_cols = [col for col in all_columns if col not in pks]
                sql2 = "UPDATE [{table}] SET {pairs} WHERE {wheres}".format(
                    table=self.name,
                    pairs=", ".join(
                        "[{}] = {}".format(col, conversions.get(col, "?"))
                        for col in set_cols
                    ),
                    wheres=" AND ".join("[{}] = ?".format(pk) for pk in pks),
                )
                queries_and_params.append(
                    (
                        sql2,
                        [record[col] for col in set_cols] + [record[pk] for pk in pks],
                    )
                )
                # We can populate .last_pk right here
                if num_records_processed == 1:
                    self.last_pk = tuple(record[pk] for pk in pks)
                    if len(self.last_pk) == 1:
                        self.last_pk = self.last_pk[0]

        else:
            or_what = ""
            if replace:
                or_what = "OR REPLACE "
            elif ignore:
                or_what = "OR IGNORE "
            sql = """
                INSERT {or_what}INTO [{table}] ({columns}) VALUES {rows};
            """.strip().format(
                or_what=or_what,
                table=self.name,
                columns=", ".join("[{}]".format(c) for c in all_columns),
                rows=", ".join(
                    "({placeholders})".format(
                        placeholders=", ".join(
                            [conversions.get(col, "?") for col in all_columns]
                        )
                    )
                    for record in chunk
                ),
            )
            flat_values = list(itertools.chain(*values))
            queries_and_params = [(sql, flat_values)]

        return queries_and_params

    def insert_chunk(
        self,
        alter,
        extracts,
        chunk,
        all_columns,
        hash_id,
        upsert,
        pk,
        conversions,
        num_records_processed,
        replace,
        ignore,
    ):
        queries_and_params = self.build_insert_queries_and_params(
            extracts,
            chunk,
            all_columns,
            hash_id,
            upsert,
            pk,
            conversions,
            num_records_processed,
            replace,
            ignore,
        )

        with self.db.conn:
            result = None
            for query, params in queries_and_params:
                try:
                    result = self.db.execute(query, params)
                except OperationalError as e:
                    if alter and (" column" in e.args[0]):
                        # Attempt to add any missing columns, then try again
                        self.add_missing_columns(chunk)
                        result = self.db.execute(query, params)
                    elif e.args[0] == "too many SQL variables":

                        first_half = chunk[: len(chunk) // 2]
                        second_half = chunk[len(chunk) // 2 :]

                        self.insert_chunk(
                            alter,
                            extracts,
                            first_half,
                            all_columns,
                            hash_id,
                            upsert,
                            pk,
                            conversions,
                            num_records_processed,
                            replace,
                            ignore,
                        )

                        self.insert_chunk(
                            alter,
                            extracts,
                            second_half,
                            all_columns,
                            hash_id,
                            upsert,
                            pk,
                            conversions,
                            num_records_processed,
                            replace,
                            ignore,
                        )

                    else:
                        raise
            if num_records_processed == 1 and not upsert:
                self.last_rowid = result.lastrowid
                self.last_pk = self.last_rowid
                # self.last_rowid will be 0 if a "INSERT OR IGNORE" happened
                if (hash_id or pk) and self.last_rowid:
                    row = list(self.rows_where("rowid = ?", [self.last_rowid]))[0]
                    if hash_id:
                        self.last_pk = row[hash_id]
                    elif isinstance(pk, str):
                        self.last_pk = row[pk]
                    else:
                        self.last_pk = tuple(row[p] for p in pk)

        return

    def insert(
        self,
        record,
        pk=DEFAULT,
        foreign_keys=DEFAULT,
        column_order=DEFAULT,
        not_null=DEFAULT,
        defaults=DEFAULT,
        hash_id=DEFAULT,
        alter=DEFAULT,
        ignore=DEFAULT,
        replace=DEFAULT,
        extracts=DEFAULT,
        conversions=DEFAULT,
        columns=DEFAULT,
    ):
        return self.insert_all(
            [record],
            pk=pk,
            foreign_keys=foreign_keys,
            column_order=column_order,
            not_null=not_null,
            defaults=defaults,
            hash_id=hash_id,
            alter=alter,
            ignore=ignore,
            replace=replace,
            extracts=extracts,
            conversions=conversions,
            columns=columns,
        )

    def insert_all(
        self,
        records,
        pk=DEFAULT,
        foreign_keys=DEFAULT,
        column_order=DEFAULT,
        not_null=DEFAULT,
        defaults=DEFAULT,
        batch_size=DEFAULT,
        hash_id=DEFAULT,
        alter=DEFAULT,
        ignore=DEFAULT,
        replace=DEFAULT,
        truncate=False,
        extracts=DEFAULT,
        conversions=DEFAULT,
        columns=DEFAULT,
        upsert=False,
    ):
        """
        Like .insert() but takes a list of records and ensures that the table
        that it creates (if table does not exist) has columns for ALL of that
        data
        """
        pk = self.value_or_default("pk", pk)
        foreign_keys = self.value_or_default("foreign_keys", foreign_keys)
        column_order = self.value_or_default("column_order", column_order)
        not_null = self.value_or_default("not_null", not_null)
        defaults = self.value_or_default("defaults", defaults)
        batch_size = self.value_or_default("batch_size", batch_size)
        hash_id = self.value_or_default("hash_id", hash_id)
        alter = self.value_or_default("alter", alter)
        ignore = self.value_or_default("ignore", ignore)
        replace = self.value_or_default("replace", replace)
        extracts = self.value_or_default("extracts", extracts)
        conversions = self.value_or_default("conversions", conversions)
        columns = self.value_or_default("columns", columns)

        if upsert and (not pk and not hash_id):
            raise PrimaryKeyRequired("upsert() requires a pk")
        assert not (hash_id and pk), "Use either pk= or hash_id="
        if hash_id:
            pk = hash_id

        assert not (
            ignore and replace
        ), "Use either ignore=True or replace=True, not both"
        all_columns = None
        first = True
        num_records_processed = 0
        # We can only handle a max of 999 variables in a SQL insert, so
        # we need to adjust the batch_size down if we have too many cols
        records = iter(records)
        # Peek at first record to count its columns:
        try:
            first_record = next(records)
        except StopIteration:
            return self  # It was an empty list
        num_columns = len(first_record.keys())
        assert (
            num_columns <= SQLITE_MAX_VARS
        ), "Rows can have a maximum of {} columns".format(SQLITE_MAX_VARS)
        batch_size = max(1, min(batch_size, SQLITE_MAX_VARS // num_columns))
        self.last_rowid = None
        self.last_pk = None
        if truncate and self.exists():
            self.db.execute("DELETE FROM [{}];".format(self.name))
        for chunk in chunks(itertools.chain([first_record], records), batch_size):
            chunk = list(chunk)
            num_records_processed += len(chunk)
            if first:
                if not self.exists():
                    # Use the first batch to derive the table names
                    column_types = suggest_column_types(chunk)
                    column_types.update(columns or {})
                    self.create(
                        column_types,
                        pk,
                        foreign_keys,
                        column_order=column_order,
                        not_null=not_null,
                        defaults=defaults,
                        hash_id=hash_id,
                        extracts=extracts,
                    )
                all_columns = set()
                for record in chunk:
                    all_columns.update(record.keys())
                all_columns = list(sorted(all_columns))
                if hash_id:
                    all_columns.insert(0, hash_id)
            else:
                all_columns += [
                    column
                    for record in chunk
                    for column in record
                    if column not in all_columns
                ]

            validate_column_names(all_columns)
            first = False

            self.insert_chunk(
                alter,
                extracts,
                chunk,
                all_columns,
                hash_id,
                upsert,
                pk,
                conversions,
                num_records_processed,
                replace,
                ignore,
            )

        return self

    def upsert(
        self,
        record,
        pk=DEFAULT,
        foreign_keys=DEFAULT,
        column_order=DEFAULT,
        not_null=DEFAULT,
        defaults=DEFAULT,
        hash_id=DEFAULT,
        alter=DEFAULT,
        extracts=DEFAULT,
        conversions=DEFAULT,
        columns=DEFAULT,
    ):
        return self.upsert_all(
            [record],
            pk=pk,
            foreign_keys=foreign_keys,
            column_order=column_order,
            not_null=not_null,
            defaults=defaults,
            hash_id=hash_id,
            alter=alter,
            extracts=extracts,
            conversions=conversions,
            columns=columns,
        )

    def upsert_all(
        self,
        records,
        pk=DEFAULT,
        foreign_keys=DEFAULT,
        column_order=DEFAULT,
        not_null=DEFAULT,
        defaults=DEFAULT,
        batch_size=DEFAULT,
        hash_id=DEFAULT,
        alter=DEFAULT,
        extracts=DEFAULT,
        conversions=DEFAULT,
        columns=DEFAULT,
    ):
        return self.insert_all(
            records,
            pk=pk,
            foreign_keys=foreign_keys,
            column_order=column_order,
            not_null=not_null,
            defaults=defaults,
            batch_size=batch_size,
            hash_id=hash_id,
            alter=alter,
            extracts=extracts,
            conversions=conversions,
            columns=columns,
            upsert=True,
        )

    def add_missing_columns(self, records):
        needed_columns = suggest_column_types(records)
        current_columns = {c.lower() for c in self.columns_dict}
        for col_name, col_type in needed_columns.items():
            if col_name.lower() not in current_columns:
                self.add_column(col_name, col_type)
        return self

    def lookup(self, column_values):
        # lookups is a dictionary - all columns will be used for a unique index
        assert isinstance(column_values, dict)
        if self.exists():
            self.add_missing_columns([column_values])
            unique_column_sets = [set(i.columns) for i in self.indexes]
            if set(column_values.keys()) not in unique_column_sets:
                self.create_index(column_values.keys(), unique=True)
            wheres = ["[{}] = ?".format(column) for column in column_values]
            rows = list(
                self.rows_where(
                    " and ".join(wheres), [value for _, value in column_values.items()]
                )
            )
            try:
                return rows[0]["id"]
            except IndexError:
                return self.insert(column_values, pk="id").last_pk
        else:
            pk = self.insert(column_values, pk="id").last_pk
            self.create_index(column_values.keys(), unique=True)
            return pk

    def m2m(
        self,
        other_table,
        record_or_iterable=None,
        pk=DEFAULT,
        lookup=None,
        m2m_table=None,
        alter=False,
    ):
        if isinstance(other_table, str):
            other_table = self.db.table(other_table, pk=pk)
        our_id = self.last_pk
        if lookup is not None:
            assert record_or_iterable is None, "Provide lookup= or record, not both"
        else:
            assert record_or_iterable is not None, "Provide lookup= or record, not both"
        tables = list(sorted([self.name, other_table.name]))
        columns = ["{}_id".format(t) for t in tables]
        if m2m_table is not None:
            m2m_table_name = m2m_table
        else:
            # Detect if there is a single, unambiguous option
            candidates = self.db.m2m_table_candidates(self.name, other_table.name)
            if len(candidates) == 1:
                m2m_table_name = candidates[0]
            elif len(candidates) > 1:
                raise NoObviousTable(
                    "No single obvious m2m table for {}, {} - use m2m_table= parameter".format(
                        self.name, other_table.name
                    )
                )
            else:
                # If not, create a new table
                m2m_table_name = m2m_table or "{}_{}".format(*tables)
        m2m_table = self.db.table(m2m_table_name, pk=columns, foreign_keys=columns)
        if lookup is None:
            # if records is only one record, put the record in a list
            records = (
                [record_or_iterable]
                if isinstance(record_or_iterable, Mapping)
                else record_or_iterable
            )
            # Ensure each record exists in other table
            for record in records:
                id = other_table.insert(
                    record, pk=pk, replace=True, alter=alter
                ).last_pk
                m2m_table.insert(
                    {
                        "{}_id".format(other_table.name): id,
                        "{}_id".format(self.name): our_id,
                    },
                    replace=True,
                )
        else:
            id = other_table.lookup(lookup)
            m2m_table.insert(
                {
                    "{}_id".format(other_table.name): id,
                    "{}_id".format(self.name): our_id,
                },
                replace=True,
            )
        return self

    def analyze_column(
        self, column, common_limit=10, value_truncate=None, total_rows=None
    ):
        db = self.db
        table = self.name
        if total_rows is None:
            total_rows = db[table].count

        def truncate(value):
            if value_truncate is None or isinstance(value, (float, int)):
                return value
            value = str(value)
            if len(value) > value_truncate:
                value = value[:value_truncate] + "..."
            return value

        num_null = db.execute(
            "select count(*) from [{}] where [{}] is null".format(table, column)
        ).fetchone()[0]
        num_blank = db.execute(
            "select count(*) from [{}] where [{}] = ''".format(table, column)
        ).fetchone()[0]
        num_distinct = db.execute(
            "select count(distinct [{}]) from [{}]".format(column, table)
        ).fetchone()[0]
        most_common = None
        least_common = None
        if num_distinct == 1:
            value = db.execute(
                "select [{}] from [{}] limit 1".format(column, table)
            ).fetchone()[0]
            most_common = [(truncate(value), total_rows)]
        elif num_distinct != total_rows:
            most_common = [
                (truncate(r[0]), r[1])
                for r in db.execute(
                    "select [{}], count(*) from [{}] group by [{}] order by count(*) desc, [{}] limit {}".format(
                        column, table, column, column, common_limit
                    )
                ).fetchall()
            ]
            most_common.sort(key=lambda p: (p[1], p[0]), reverse=True)
            if num_distinct <= common_limit:
                # No need to run the query if it will just return the results in revers order
                least_common = None
            else:
                least_common = [
                    (truncate(r[0]), r[1])
                    for r in db.execute(
                        "select [{}], count(*) from [{}] group by [{}] order by count(*), [{}] desc limit {}".format(
                            column, table, column, column, common_limit
                        )
                    ).fetchall()
                ]
                least_common.sort(key=lambda p: (p[1], p[0]))
        return ColumnDetails(
            self.name,
            column,
            total_rows,
            num_null,
            num_blank,
            num_distinct,
            most_common,
            least_common,
        )


class View(Queryable):
    def exists(self):
        return True

    def __repr__(self):
        return "<View {} ({})>".format(
            self.name, ", ".join(c.name for c in self.columns)
        )

    def drop(self):
        self.db.execute("DROP VIEW [{}]".format(self.name))


def chunks(sequence, size):
    iterator = iter(sequence)
    for item in iterator:
        yield itertools.chain([item], itertools.islice(iterator, size - 1))


def jsonify_if_needed(value):
    if isinstance(value, decimal.Decimal):
        return float(value)
    if isinstance(value, (dict, list, tuple)):
        return json.dumps(value, default=repr)
    elif isinstance(value, (datetime.time, datetime.date, datetime.datetime)):
        return value.isoformat()
    elif isinstance(value, uuid.UUID):
        return str(value)
    else:
        return value


def _hash(record):
    return hashlib.sha1(
        json.dumps(record, separators=(",", ":"), sort_keys=True, default=repr).encode(
            "utf8"
        )
    ).hexdigest()


def resolve_extracts(extracts):
    if extracts is None:
        extracts = {}
    if isinstance(extracts, (list, tuple)):
        extracts = {item: item for item in extracts}
    return extracts


def validate_column_names(columns):
    # Validate no columns contain '[' or ']' - #86
    for column in columns:
        assert (
            "[" not in column and "]" not in column
        ), "'[' and ']' cannot be used in column names"