Given this schema: sql CREATE TABLE departments ( campus_name TEXT NOT NULL, dept_code TEXT NOT NULL, dept_name TEXT, PRIMARY KEY (campus_name, dept_code) ); CREATE TABLE courses ( course_code TEXT PRIMARY KEY, course_name TEXT, campus_name TEXT NOT NULL, dept_code TEXT NOT NULL, FOREIGN KEY (campus_name, dept_code) REFERENCES departments(campus_name, dept_code) ); The output of db["courses"].foreign_keys right now is: [ForeignKey(table='courses', column='campus_name', other_table='departments', other_column='campus_name'), ForeignKey(table='courses', column='dept_code', other_table='departments', other_column='dept_code')] Which suggests two normal foreign keys, not one compound foreign key.

Version 3.33

sqlite-utils query library.db 'select asin from audible' --fmt plain --no-headers | head -3 asin 0062804006 0062891421

sqlite supports these, and it would be quite nice to be able to add them with sqlite-utils.

https://www.sqlite.org/foreignkeys.html#fk_actions

https://github.com/simonw/sqlite-utils/pull/498

:books: Documentation preview :books:: https://sqlite-utils--499.org.readthedocs.build/en/499/

alternatively, self.disable_fts()

I'm trying to figure out a way to get the table.extract() method to complete successfully -- I'm not sure if maybe the cause (and a possible solution) of this on Ubuntu Server 22.04 is to adjust some of the PRAGMA values within SQLite itself ... on another Linux system (PopOS), using this method on this same database appears to work just fine.

Here's the bit that's causing the error, and the resulting error output: ```python

combine these columns into 1 table "bib_properties" :

best_title

bib_level_code

mat_type

material_code

best_author

db["circ_trans"].extract( ["best_title", "bib_level_code", "mat_type", "material_code", "best_author"], table="bib_properties", fk_column="bib_properties_id" )

db["circ_trans"].extract( ["call_number"], table="call_number", fk_column="call_number_id", rename={"call_number": "value"} ) ```

```python

OperationalError Traceback (most recent call last) Input In [17], in <cell line: 7>() 1 # combine these columns into 1 table "bib_properties" : 2 # best_title 3 # bib_level_code 4 # mat_type 5 # material_code 6 # best_author ----> 7 db["circ_trans"].extract( 8 ["best_title", "bib_level_code", "mat_type", "material_code", "best_author"], 9 table="bib_properties", 10 fk_column="bib_properties_id" 11 ) 13 db["circ_trans"].extract( 14 ["call_number"], 15 table="call_number", 16 fk_column="call_number_id", 17 rename={"call_number": "value"} 18 )

File ~/jupyter/venv/lib/python3.10/site-packages/sqlite_utils/db.py:1764, in Table.extract(self, columns, table, fk_column, rename) 1761 column_order.append(c.name) 1763 # Drop the unnecessary columns and rename lookup column -> 1764 self.transform( 1765 drop=set(columns), 1766 rename={magic_lookup_column: fk_column}, 1767 column_order=column_order, 1768 ) 1770 # And add the foreign key constraint 1771 self.add_foreign_key(fk_column, table, "id")

File ~/jupyter/venv/lib/python3.10/site-packages/sqlite_utils/db.py:1526, in Table.transform(self, types, rename, drop, pk, not_null, defaults, drop_foreign_keys, column_order) 1524 with self.db.conn: 1525 for sql in sqls: -> 1526 self.db.execute(sql) 1527 # Run the foreign_key_check before we commit 1528 if pragma_foreign_keys_was_on:

File ~/jupyter/venv/lib/python3.10/site-packages/sqlite_utils/db.py:465, in Database.execute(self, sql, parameters) 463 return self.conn.execute(sql, parameters) 464 else: --> 465 return self.conn.execute(sql)

OperationalError: database or disk is full ```

This database is about 17G in total size, so I'm assuming the error is coming from the vacuum ... where i'm assuming it's maybe trying to do the temp storage in a location that doesn't have sufficient room. The disk space is more than ample on the host in question (1.8T is free in the directory where the sqlite db resides) The /tmp directory however is limited on a smaller disk associated with the OS

I'm trying to think if there's a way to set the PRAGMA temp_store or maybe if it's temp_store_directory that I'm after ... to use the same local directory of where the file is located (maybe this is a property of the version of sqlite on the system?)

```python

SET the temp file store to be a file ...

print(db.execute('PRAGMA temp_store').fetchall()) print(db.execute('PRAGMA temp_store=FILE').fetchall())

print(db.execute('PRAGMA temp_store').fetchall())

the users home directory ...

print(db.execute("PRAGMA temp_store_directory='/home/plchuser/'").fetchall()) print(db.execute("PRAGMA sqlite3_temp_directory='/home/plchuser/'").fetchall())

print(db.execute("PRAGMA temp_store_directory").fetchall()) print(db.execute("PRAGMA sqlite3_temp_directory").fetchall()) text [(1,)] [] [(1,)] [] [] [('/home/plchuser/',)] [] ```

Here's the docs on the Temporary File Storage Locations https://www.sqlite.org/tempfiles.html

Would it be worthwhile for the sqlite-utils command-line tool to grow features to efficiently directly interact with gzipped CSV data?

Maybe add --gz options to both insert and to the various commands that output query results.

Can show people how to open this up in Binder.

I'm getting this error when running:

sqlite-utils enable-wal beta.db

OperationalError: cannot change into wal mode from within a transaction

I'm worried that maybe that's because of this new code from #152:

https://github.com/simonw/sqlite-utils/blob/deb2eb013ff85bbc828ebc244a9654f0d9c3139e/sqlite_utils/db.py#L128-L129

Hi! I did not find any mention on the library about ON DELETE and ON UPDATE actions for foreign keys. Are those expected to be implemented? If not, it would be a nice thing to include!