From 64799df78b14a12084d1def91c561abdcbcd8773 Mon Sep 17 00:00:00 2001
From: Simon Willison <swillison@gmail.com>
Date: Tue, 8 Sep 2020 15:18:12 -0700
Subject: [PATCH] .optimize() no longer cleans up _docsize

This isn't necessary any more since the new .rebuild_fts()
method can achieve the same thing.

Refs #155, #153
---
 docs/python-api.rst    |  8 ++++----
 sqlite_utils/db.py     | 14 --------------
 tests/test_fts.py      | 18 ++++++++++++++++++
 tests/test_optimize.py | 20 --------------------
 4 files changed, 22 insertions(+), 38 deletions(-)
 delete mode 100644 tests/test_optimize.py

diff --git a/docs/python-api.rst b/docs/python-api.rst
index d29133e..2564376 100644
--- a/docs/python-api.rst
+++ b/docs/python-api.rst
@@ -1228,6 +1228,10 @@ This method can be called on a table that has been configured for full-text sear
 
     db["dogs_fts"].rebuild_fts()
 
+This runs the following SQL::
+
+    INSERT INTO dogs_fts (dogs_fts) VALUES ("rebuild");
+
 Optimizing a full-text search table
 ===================================
 
@@ -1240,10 +1244,6 @@ Once you have populated a FTS table you can optimize it to dramatically reduce i
 This runs the following SQL::
 
     INSERT INTO dogs_fts (dogs_fts) VALUES ("optimize");
-    DELETE FROM [dogs_fts_docsize] WHERE id NOT IN (
-        SELECT rowid FROM [dogs_fts]);
-
-That ``DELETE`` statement cleans up rows that may have been created by `an obscure bug <https://github.com/simonw/sqlite-utils/issues/153>`__ in previous versions of ``sqlite-utils``.
 
 Creating indexes
 ================
diff --git a/sqlite_utils/db.py b/sqlite_utils/db.py
index e077231..c03530f 100644
--- a/sqlite_utils/db.py
+++ b/sqlite_utils/db.py
@@ -951,20 +951,6 @@ class Table(Queryable):
                     table=fts_table
                 )
             )
-            self.db.execute(
-                textwrap.dedent(
-                    """
-                DELETE FROM [{table}_docsize] WHERE {column} NOT IN (
-                    SELECT rowid FROM [{table}]);
-            """
-                )
-                .strip()
-                .format(
-                    # FTS5 uses 'id' but FTS4 uses 'docid'
-                    column=self.db["{}_docsize".format(fts_table)].columns[0].name,
-                    table=fts_table,
-                )
-            )
         return self
 
     def search(self, q):
diff --git a/tests/test_fts.py b/tests/test_fts.py
index e0b5475..fb94a2e 100644
--- a/tests/test_fts.py
+++ b/tests/test_fts.py
@@ -1,4 +1,5 @@
 import pytest
+from sqlite_utils import Database
 from sqlite_utils.utils import sqlite3
 
 search_records = [
@@ -199,3 +200,20 @@ def test_rebuild_fts_invalid(fresh_db, invalid_table):
     # Raise OperationalError on invalid table
     with pytest.raises(sqlite3.OperationalError):
         fresh_db[invalid_table].rebuild_fts()
+
+
+@pytest.mark.parametrize("fts_version", ["FTS4", "FTS5"])
+def test_rebuild_removes_junk_docsize_rows(tmpdir, fts_version):
+    # Recreating https://github.com/simonw/sqlite-utils/issues/149
+    path = tmpdir / "test.db"
+    db = Database(str(path), recursive_triggers=False)
+    licenses = [{"key": "apache2", "name": "Apache 2"}, {"key": "bsd", "name": "BSD"}]
+    db["licenses"].insert_all(licenses, pk="key", replace=True)
+    db["licenses"].enable_fts(["name"], create_triggers=True, fts_version=fts_version)
+    assert db["licenses_fts_docsize"].count == 2
+    # Bug: insert with replace increases the number of rows in _docsize:
+    db["licenses"].insert_all(licenses, pk="key", replace=True)
+    assert db["licenses_fts_docsize"].count == 4
+    # rebuild should fix this:
+    db["licenses_fts"].rebuild_fts()
+    assert db["licenses_fts_docsize"].count == 2
diff --git a/tests/test_optimize.py b/tests/test_optimize.py
deleted file mode 100644
index 302532c..0000000
--- a/tests/test_optimize.py
+++ /dev/null
@@ -1,20 +0,0 @@
-import pytest
-from sqlite_utils import Database
-import sqlite3
-
-
-@pytest.mark.parametrize("fts_version", ["FTS4", "FTS5"])
-def test_optimizes_removes_junk_docsize_rows(tmpdir, fts_version):
-    # Recreating https://github.com/simonw/sqlite-utils/issues/149
-    path = tmpdir / "test.db"
-    db = Database(str(path), recursive_triggers=False)
-    licenses = [{"key": "apache2", "name": "Apache 2"}, {"key": "bsd", "name": "BSD"}]
-    db["licenses"].insert_all(licenses, pk="key", replace=True)
-    db["licenses"].enable_fts(["name"], create_triggers=True, fts_version=fts_version)
-    assert db["licenses_fts_docsize"].count == 2
-    # Bug: insert with replace increases the number of rows in _docsize:
-    db["licenses"].insert_all(licenses, pk="key", replace=True)
-    assert db["licenses_fts_docsize"].count == 4
-    # Optimize should fix this:
-    db["licenses_fts"].optimize()
-    assert db["licenses_fts_docsize"].count == 2