issue_comments

where did you see the limit as 999? I believe the limit has been 32766 for quite some time. If you could detect which one this could speed up batch insert of some types of data significantly

Here's the diff where I tried to use .executemany() and ran into the lastrowid problem: diff diff --git a/sqlite_utils/db.py b/sqlite_utils/db.py index ef55976..7f85759 100644 --- a/sqlite_utils/db.py +++ b/sqlite_utils/db.py @@ -881,13 +881,10 @@ class Table: or_what=or_what, table=self.name, columns=", ".join("[{}]".format(c) for c in all_columns), - rows=", ".join( - """ + rows=""" ({placeholders}) """.format( - placeholders=", ".join(["?"] * len(all_columns)) - ) - for record in chunk + placeholders=", ".join(["?"] * len(all_columns)) ), ) values = [] @@ -902,15 +899,15 @@ class Table: extract_table = extracts[key] value = self.db[extract_table].lookup({"value": value}) record_values.append(value) - values.extend(record_values) + values.append(record_values) with self.db.conn: try: - result = self.db.conn.execute(sql, values) + result = self.db.conn.executemany(sql, values) except sqlite3.OperationalError as e: if alter and (" has no column " in e.args[0]): # Attempt to add any missing columns, then try again self.add_missing_columns(chunk) - result = self.db.conn.execute(sql, values) + result = self.db.conn.executemany(sql, values) else: raise self.last_rowid = result.lastrowid

The problem with .executemany() is it breaks lastrowid:

This read-only attribute provides the rowid of the last modified row. It is only set if you issued an INSERT or a REPLACE statement using the execute() method. For operations other than INSERT or REPLACE or when executemany() is called, lastrowid is set to None.

So I think I need to continue to use my existing way of executing bulk inserts (with a giant repeated INSERT INTO ... VALUES block) but ensure that I calculate the chunk size such that I don't ever try to pass more than 999 values at once.

I think the fix here is for me to switch to using executemany() - example from the Python docs: https://docs.python.org/3/library/sqlite3.html python purchases = [('2006-03-28', 'BUY', 'IBM', 1000, 45.00), ('2006-04-05', 'BUY', 'MSFT', 1000, 72.00), ('2006-04-06', 'SELL', 'IBM', 500, 53.00), ] c.executemany('INSERT INTO stocks VALUES (?,?,?,?,?)', purchases)