From 736dca6289fd087ba8d810456e4c7f9f49a584e5 Mon Sep 17 00:00:00 2001
From: Simon Willison <swillison@gmail.com>
Date: Sat, 12 Dec 2020 21:28:43 -0800
Subject: [PATCH] Much improved design for CLI output of analyze-tables

---
 sqlite_utils/cli.py          | 41 +++++++++++++++++++++++++++++++++++-
 sqlite_utils/db.py           |  2 +-
 tests/test_analyze_tables.py | 39 +++++++++++++++++++++++++++++-----
 3 files changed, 75 insertions(+), 7 deletions(-)

diff --git a/sqlite_utils/cli.py b/sqlite_utils/cli.py
index 02111a9..81454dc 100644
--- a/sqlite_utils/cli.py
+++ b/sqlite_utils/cli.py
@@ -7,6 +7,7 @@ import hashlib
 import pathlib
 import sqlite_utils
 from sqlite_utils.db import AlterError
+import textwrap
 import itertools
 import json
 import os
@@ -1399,8 +1400,46 @@ def analyze_tables(
             db["_analyze_tables_"].insert(
                 column_details._asdict(), pk=("table", "column"), replace=True
             )
+        most_common_rendered = _render_common(
+            "\n\n  Most common:", column_details.most_common
+        )
+        least_common_rendered = _render_common(
+            "\n\n  Least common:", column_details.least_common
+        )
+        details = (
+            (
+                textwrap.dedent(
+                    """
+        {table}.{column}: ({i}/{total})
 
-        click.echo("{}/{}: {}".format(i + 1, len(todo), column_details))
+          Total rows: {total_rows}
+          Null rows: {num_null}
+          Blank rows: {num_blank}
+
+          Distinct values: {num_distinct}{most_common_rendered}{least_common_rendered}
+        """
+                )
+                .strip()
+                .format(
+                    i=i + 1,
+                    total=len(todo),
+                    most_common_rendered=most_common_rendered,
+                    least_common_rendered=least_common_rendered,
+                    **column_details._asdict()
+                )
+            )
+            + "\n"
+        )
+        click.echo(details)
+
+
+def _render_common(title, values):
+    if values is None:
+        return ""
+    lines = [title]
+    for value, count in values:
+        lines.append("    {}: {}".format(value, count))
+    return "\n".join(lines)
 
 
 FILE_COLUMNS = {
diff --git a/sqlite_utils/db.py b/sqlite_utils/db.py
index 1974860..9f2653c 100644
--- a/sqlite_utils/db.py
+++ b/sqlite_utils/db.py
@@ -1944,7 +1944,7 @@ class Table(Queryable):
         return self
 
     def analyze_column(
-        self, column, common_limit=10, value_truncate=100, total_rows=None
+        self, column, common_limit=10, value_truncate=80, total_rows=None
     ):
         db = self.db
         table = self.name
diff --git a/tests/test_analyze_tables.py b/tests/test_analyze_tables.py
index 5aa8269..17ec966 100644
--- a/tests/test_analyze_tables.py
+++ b/tests/test_analyze_tables.py
@@ -89,12 +89,41 @@ def test_analyze_table(db_to_analyze_path):
     result = CliRunner().invoke(cli.cli, ["analyze-tables", db_to_analyze_path])
     assert (
         result.output.strip()
-        == textwrap.dedent(
+        == (
             """
-    1/3: ColumnDetails(table='stuff', column='id', total_rows=8, num_null=0, num_blank=0, num_distinct=8, most_common=None, least_common=None)
-    2/3: ColumnDetails(table='stuff', column='owner', total_rows=8, num_null=0, num_blank=0, num_distinct=4, most_common=[('Joan', 3), ('Terryterryterry', 2), ('Kumar', 2), ('Anne', 1)], least_common=None)
-    3/3: ColumnDetails(table='stuff', column='size', total_rows=8, num_null=0, num_blank=0, num_distinct=2, most_common=[(5, 5), (4, 3)], least_common=None)
-    """
+stuff.id: (1/3)
+
+  Total rows: 8
+  Null rows: 0
+  Blank rows: 0
+
+  Distinct values: 8
+
+stuff.owner: (2/3)
+
+  Total rows: 8
+  Null rows: 0
+  Blank rows: 0
+
+  Distinct values: 4
+
+  Most common:
+    Joan: 3
+    Terryterryterry: 2
+    Kumar: 2
+    Anne: 1
+
+stuff.size: (3/3)
+
+  Total rows: 8
+  Null rows: 0
+  Blank rows: 0
+
+  Distinct values: 2
+
+  Most common:
+    5: 5
+    4: 3"""
         ).strip()
     )