issues

https://dev.to/hugovk/help-test-python-312-beta-1508/

:books: Documentation preview :books:: https://sqlite-utils--591.org.readthedocs.build/en/591/

I used a custom SQL function in a migration script and then realized that it should be de-registered before the end of the script to avoid leaking into the calling code.

Extremely detailed story of how I got to this point:

• https://github.com/simonw/llm/issues/162

Steps to reproduce (only if that pragma is on though): bash python -c ' import sqlite_utils db = sqlite_utils.Database(memory=True) db.execute(""" CREATE TABLE "logs" ( [id] INTEGER PRIMARY KEY, [model] TEXT, [prompt] TEXT, [system] TEXT, [prompt_json] TEXT, [options_json] TEXT, [response] TEXT, [response_json] TEXT, [reply_to_id] INTEGER, [chat_id] INTEGER REFERENCES [log]([id]), [duration_ms] INTEGER, [datetime_utc] TEXT ); """) db["logs"].add_foreign_key("reply_to_id", "logs", "id") ' This succeeds in some environments, fails in others.

I got this error trying to drop a column from a table that was part of a SQL view:

error in view plugins: no such table: main.pypi_releases

Upon further investigation I found that this pattern seemed to fix it: python def transform_the_table(conn): # Run this in a transaction: with conn: # We have to read all the views first, because we need to drop and recreate them db = sqlite_utils.Database(conn) views = {v.name: v.schema for v in db.views if table.lower() in v.schema.lower()} for view in views.keys(): db[view].drop() db[table].transform( types=types, rename=rename, drop=drop, column_order=[p[0] for p in order_pairs], ) # Now recreate the views for name, schema in views.items(): db.create_view(name, schema) So grab a copy of any view that might reference this table, start a transaction, drop those views, run the transform, recreate the views again.

I wonder if this should become an option in sqlite-utils? Maybe a recreate_views=True argument for table.tranform(...)? Should it be opt-in or opt-out?

Originally posted by @simonw in https://github.com/simonw/datasette-edit-schema/issues/35#issuecomment-1683370548

Splitting off an issue for prepare_connection() since Alex got the PR in seconds before I shipped 3.34!

Originally posted by @simonw in https://github.com/simonw/sqlite-utils/issues/567#issuecomment-1646686424

I'm going to start by adding the register_command hook using the exact same pattern as Datasette and LLM.

Originally posted by @simonw in https://github.com/simonw/sqlite-utils/issues/567#issuecomment-1646643450

https://github.com/simonw/sqlite-utils/actions/runs/5039064797/jobs/9036965488

Traceback (most recent call last): File "/opt/hostedtoolcache/Python/3.7.16/x64/bin/flake8", line 8, in <module> sys.exit(main()) File "/opt/hostedtoolcache/Python/3.7.16/x64/lib/python3.7/site-packages/flake8/main/cli.py", line 22, in main app.run(argv) File "/opt/hostedtoolcache/Python/3.7.16/x64/lib/python3.7/site-packages/flake8/main/application.py", line 363, in run self._run(argv) File "/opt/hostedtoolcache/Python/3.7.16/x64/lib/python3.7/site-packages/flake8/main/application.py", line 350, in _run self.initialize(argv) File "/opt/hostedtoolcache/Python/3.7.16/x64/lib/python3.7/site-packages/flake8/main/application.py", line 330, in initialize self.find_plugins(config_finder) File "/opt/hostedtoolcache/Python/3.7.16/x64/lib/python3.7/site-packages/flake8/main/application.py", line 153, in find_plugins self.check_plugins = plugin_manager.Checkers(local_plugins.extension) File "/opt/hostedtoolcache/Python/3.7.16/x64/lib/python3.7/site-packages/flake8/plugins/manager.py", line 357, in __init__ self.namespace, local_plugins=local_plugins File "/opt/hostedtoolcache/Python/3.7.16/x64/lib/python3.7/site-packages/flake8/plugins/manager.py", line 238, in __init__ self._load_entrypoint_plugins() File "/opt/hostedtoolcache/Python/3.7.16/x64/lib/python3.7/site-packages/flake8/plugins/manager.py", line 254, in _load_entrypoint_plugins eps = importlib_metadata.entry_points().get(self.namespace, ()) AttributeError: 'EntryPoints' object has no attribute 'get' Error: Process completed with exit code 1.

Refs: - #545

:books: Documentation preview :books:: https://sqlite-utils--549.org.readthedocs.build/en/549/

I got this error:

File "/Users/simon/Dropbox/Development/openai-to-sqlite/openai_to_sqlite/cli.py", line 27, in embeddings rows, _ = rows_from_file(input) ^^^^^^^^^^^^^^^^^^^^^ File "/Users/simon/.local/share/virtualenvs/openai-to-sqlite-jt4obeb2/lib/python3.11/site-packages/sqlite_utils/utils.py", line 305, in rows_from_file first_bytes = buffered.peek(2048).strip() ^^^^^^^^^^^^^^^^^^^ From this code: ```python

@cli.command() @click.argument( "db_path", type=click.Path(file_okay=True, dir_okay=False, allow_dash=False), ) @click.option( "-i", "--input", type=click.File("r"), default="-", ) def embeddings(db_path, input): "Store embeddings for one or more text documents" click.echo("Here is some output") db = sqlite_utils.Database(db_path) rows, _ = rows_from_file(input) print(list(rows)) `` The error went away when I changed it totype=click.File("rb")`.

This should either be called out in the documentation or rows_from_file() should be fixed to handle text-mode files in addition to binary files.

https://github.com/simonw/sqlite-utils/actions/runs/3472012235 failed on Python 3.11:

Truncated output: sqlite_utils/db.py:2467: note: PEP 484 prohibits implicit Optional. Accordingly, mypy has changed its default to no_implicit_optional=True sqlite_utils/db.py:2467: note: Use https://github.com/hauntsaninja/no_implicit_optional to automatically upgrade your codebase sqlite_utils/db.py:2530: error: Incompatible default for argument "where" (default has type "None", argument has type "str") [assignment] sqlite_utils/db.py:2530: note: PEP 484 prohibits implicit Optional. Accordingly, mypy has changed its default to no_implicit_optional=True sqlite_utils/db.py:2530: note: Use https://github.com/hauntsaninja/no_implicit_optional to automatically upgrade your codebase sqlite_utils/db.py:2658: error: Argument 1 to "count_where" of "Queryable" has incompatible type "Optional[str]"; expected "str" [arg-type] Found 23 errors in 1 file (checked 51 source files) Best look at https://github.com/hauntsaninja/no_implicit_optional

In building this Datasette feature on top of sqlite-utils I thought it might be useful to expose the number of rows that had been affected by a bulk insert or update - the cursor.rowcount:

• https://github.com/simonw/datasette/issues/1866

This isn't currently exposed by sqlite-utils.

Added in: - #483

I don't know how to fix this in Sphinx: I'm getting this: https://sqlite-utils.datasette.io/en/latest/cli.html#cli-install

The insert –convert and query –functions options

But I want it to display insert --convert and not insert –convert there.

Here's the code: https://github.com/simonw/sqlite-utils/blob/85247038f70d7eb2f3e272cfeaa4c44459cafba8/docs/cli.rst#L2125

This can be a documented wrapper around with db.conn:.

https://sqlite-utils.datasette.io/en/stable/cli-reference.html#insert

It looks like Python keywords are being incorrectly highlighted here.

I've used it in a couple of external places now:

• https://github.com/simonw/datasette-socrata/blob/32fb256a461bf0e790eca10bdc7dd9d96c20f7c4/datasette_socrata/init.py#L264-L280
• https://github.com/simonw/datasette-lite/blob/caa8eade10f0321c64f9f65c4561186f02d57c5b/webworker.js#L55-L64

Refs: - https://github.com/simonw/datasette-lite/issues/32

I built this for shapefile-to-sqlite but it would be useful in sqlite-utils too:

https://github.com/simonw/shapefile-to-sqlite/blob/e754d0747ca2facf9a7433e2d5d15a6a37a9cf6e/shapefile_to_sqlite/utils.py#L16-L19

python SPATIALITE_PATHS = ( "/usr/lib/x86_64-linux-gnu/mod_spatialite.so", "/usr/local/lib/mod_spatialite.dylib", )

https://github.com/simonw/shapefile-to-sqlite/blob/e754d0747ca2facf9a7433e2d5d15a6a37a9cf6e/shapefile_to_sqlite/utils.py#L105-L109

python def find_spatialite(): for path in SPATIALITE_PATHS: if os.path.exists(path): return path return None

In conjunction with #135 - this would do the same thing as --load-extension=path-to-spatialite (see #134)

The .executemany() method is a very efficient way to execute the same SQL query against a huge list of parameters.

sqlite-utils insert supports a bunch of ways of loading a list of dictionaries - from CSV, TSV, JSON, newline JSON and more thanks to: - #361

What if you could load a list of dictionaries and provide a SQL query with :named parameters that correspond to keys in those dictionaries instead?

This would need to be a new command - I thought about adding a --sql option to insert but that doesn't make sense as that command already requires a table name.

Refs #375

Still needs:

• [x] Refactor @insert_upsert_options so that it doesn't duplicate @import_options
• [x] Tests
• [x] Documentation
• [x] Try it against a really big file

https://sqlite-utils.datasette.io/en/stable/cli.html#table-formatted-output currently cheats and tells the user to run --help - can fix this using cog.

Add this to sqlite_utils/cli.py:

python if __name__ == "__main__": cli() Now the tool can be run using python -m sqlite_utils.cli --help

Originally posted by @simonw in https://github.com/simonw/sqlite-utils/issues/364#issuecomment-1008214998

I'm writing code for https://github.com/simonw/git-history/issues/33 that creates a table inside a loop:

python item_pk = db[item_table].lookup( {"_item_id": item_id}, item_to_insert, column_order=("_id", "_item_id"), pk="_id", ) I need to look things up by _item_id on this table, which means I need an index on that column (the table can get very big).

But there's no mechanism in SQLite utils to detect if the table was created for the first time and add an index to it. And I don't want to run CREATE INDEX IF NOT EXISTS every time through the loop.

This should work like the foreign_keys= mechanism.

Inspired by work on git-history where I used this pattern: ```python column_name_to_id = {}

def column_id(column):
    if column not in column_name_to_id:
        id = db["columns"].lookup(
            {"namespace": namespace_id, "name": column},
            foreign_keys=(("namespace", "namespaces", "id"),),
        )
        column_name_to_id[column] = id
    return column_name_to_id[column]

`` If you're going to be doing a large number oftable.lookup(...)` calls and you know that no other script will be modifying the database at the same time you can presumably get a big speedup using a Python in-memory cache - maybe even a LRU one to avoid memory bloat.

I had to do this with a jq recipe today: https://til.simonwillison.net/cloudrun/tailing-cloud-run-request-logs

cat log.json | jq -c '[leaf_paths as $path | { "key": $path | join("_"), "value": getpath($path) }] | from_entries' \ | sqlite-utils insert /tmp/logs.db logs - --nl --alter --batch-size 1 That was to turn something like this: json { "httpRequest": { "latency": "0.112114537s", "requestMethod": "GET", "requestSize": "534", "status": 200, }, "insertId": "6111722f000b5b4c4d4071e2", "labels": { "service": "datasette-io" } } Into this instead: json { "httpRequest_latency": "0.112114537s", "httpRequest_requestMethod": "GET", "httpRequest_requestSize": "534", "httpRequest_status": 200, "insertId": "6111722f000b5b4c4d4071e2", "labels_service": "datasette-io" } I have to do this often enough that I think it should be an option, --flatten - so I can do this instead: cat log.json | sqlite-utils insert /tmp/logs.db logs - --flatten

A good starting point would be adding type information to the members of these named tuples and the introspection methods that return them:

https://github.com/simonw/sqlite-utils/blob/9dff7a38831d471b1dff16d40d89eb5c3b4e84d6/sqlite_utils/db.py#L51-L75

For applying the conversion to a subset of rows selected using the where clause.

Should also take optional arguments, as seen in db["dogs"].delete_where("age < ?", [3]).

Follows #302 and #251. This was originally https://github.com/simonw/sqlite-transform/issues/9

I often find sqlite-utils insert ... --csv creates a first column with a weird character at the start of it - which it turns out is the UTF-8 BOM. Fix that.

While working on search (#192) I've spotted a few small changes I would like to make that would break backwards compatibility in minor ways, hence requiring a 3.x release.

db[table].search() - I would like this to default to sorting by rank

Also I'd like to free up the -c and -f options for other purposes from the standard output formats here:

https://github.com/simonw/sqlite-utils/blob/43eae8b193d362f2b292df73e087ed6f10838144/sqlite_utils/cli.py#L48-L58

I'd like -f to be used to indicate a full-text search column during an insert and -c to indicate a column (so you can specify which columns you want to output).

Refs #192. Still needs tests.

@simonw any thoughts on allow extracts to specify the lookup column name? If I'm understanding the documentation right, .lookup() allows you to define the "value" column (the documentation uses name), but when you use extracts keyword as part of .insert(), .upsert() etc. the lookup must be done against a column named "value". I have an existing lookup table that I've populated with columns "id" and "name" as opposed to "id" and "value", and seems I can't use extracts=, unless I'm missing something...

Initial thought on how to do this would be to allow the dictionary value to be a tuple of table name column pair... so: table = db.table("trees", extracts={"species_id": ("Species", "name"})

I haven't dug too much into the existing code yet, but does this make sense? Worth doing?

Originally posted by @chrishas35 in https://github.com/simonw/sqlite-utils/issues/46#issuecomment-592999503

Useful if you add one but make a mistake and need to undo it without recreating the database from scratch.

Originally posted by @simonw in https://github.com/simonw/sqlite-utils/pull/118#issuecomment-655283393

We should put some thought into how this library supports and encourages smart use of transactions.

Need to update docs for .transform() now that column_order= is available. Originally posted by @simonw in https://github.com/simonw/sqlite-utils/pull/174#discussion_r494403327

Maybe also add this as an option to sqlite-utils transform - since reordering columns is actually a pretty nice capability.

The second challenge here is cleaning up all of those junk rows in existing *_fts_docsize tables. Doing that just to the demo database from https://github-to-sqlite.dogsheep.net/github.db dropped its size from 22MB to 16MB! Here's the SQL: sql DELETE FROM [licenses_fts_docsize] WHERE id NOT IN ( SELECT rowid FROM [licenses_fts]); I can do that as part of the existing table.optimize() method, which optimizes FTS tables. Originally posted by @simonw in https://github.com/simonw/sqlite-utils/issues/149#issuecomment-688501064

FTS5 supports things like porter stemming using a tokenize= option:

https://www.sqlite.org/fts5.html#tokenizers

Something like this in code: CREATE VIRTUAL TABLE [{table}_fts] USING {fts_version} ( {columns}, tokenize='porter', content=[{table}] ); I tried this out just now and it worked exactly as expected.

So... db[table].enable_fts(...) should accept a 'tokenize= argument, and sqlite-utils enable-fts ... should support a --tokenize option.

Inserting files by piping them in should work - but since a filename cannot be derived this will need a --name blah.gif option.

cat blah.gif | sqlite-utils insert-files files.db files - --name=blah.gif

Originally posted by @simonw in https://github.com/simonw/sqlite-utils/issues/122#issuecomment-664128071

It turns out SQLite supports composite foreign keys: https://www.sqlite.org/foreignkeys.html#fk_composite

Their example looks like this: ```sql CREATE TABLE album( albumartist TEXT, albumname TEXT, albumcover BINARY, PRIMARY KEY(albumartist, albumname) );

CREATE TABLE song( songid INTEGER, songartist TEXT, songalbum TEXT, songname TEXT, FOREIGN KEY(songartist, songalbum) REFERENCES album(albumartist, albumname) ); ```

Here's what that looks like in sqlite-utils:

``` In [1]: import sqlite_utils

In [2]: import sqlite3

In [3]: conn = sqlite3.connect(":memory:")

In [4]: conn
Out[4]: <sqlite3.Connection at 0x1087186c0>

In [5]: conn.executescript(""" ...: CREATE TABLE album( ...: albumartist TEXT, ...: albumname TEXT, ...: albumcover BINARY, ...: PRIMARY KEY(albumartist, albumname) ...: ); ...:
...: CREATE TABLE song( ...: songid INTEGER, ...: songartist TEXT, ...: songalbum TEXT, ...: songname TEXT, ...: FOREIGN KEY(songartist, songalbum) REFERENCES album(albumartist, albumname) ...: ); ...: """)
Out[5]: <sqlite3.Cursor at 0x1088def10>

In [6]: db = sqlite_utils.Database(conn)

In [7]: db.tables
Out[7]: [<Table album (albumartist, albumname, albumcover)>, <Table song (songid, songartist, songalbum, songname)>] In [8]: db.tables[0].foreign_keys Out[8]: [] In [9]: db.tables[1].foreign_keys Out[9]: [ForeignKey(table='song', column='songartist', other_table='album', other_column='albumartist'), ForeignKey(table='song', column='songalbum', other_table='album', other_column='albumname')] ``` The table appears to have two separate foreign keys, when actually it has a single compound composite foreign key.

As seen in https://github.com/dogsheep/github-to-sqlite/issues/27#issuecomment-614843406 - there's a problem where you insert a record with a None value for a column and that column is created as TEXT - but actually you intended it to be an INT (as later examples will demonstrate).

Some kind of mechanism for over-riding the detected types of columns would be useful here.

Relates to #42 and #44. I want the ability to extract values out into lookup tables during bulk insert/upsert operations.

db.insert_all(rows, extracts=["species"])

• creates species table for values in the species column

db.insert_all(rows, extracts={"species": "Species"})

• as above but the new table is called Species.

https://travis-ci.com/simonw/sqlite-utils/jobs/268436167

This is a recent change: 2 months ago they worked fine.

I'm not sure what changed here. Maybe something to do with https://launchpad.net/~jonathonf/+archive/ubuntu/backports ?

I spotted a table which was created once and then had columns added to it and the formatted SQL looks like this:

sql CREATE TABLE [records] ( [type] TEXT, [sourceName] TEXT, [sourceVersion] TEXT, [unit] TEXT, [creationDate] TEXT, [startDate] TEXT, [endDate] TEXT, [value] TEXT, [metadata_Health Mate App Version] TEXT, [metadata_Withings User Identifier] TEXT, [metadata_Modified Date] TEXT, [metadata_Withings Link] TEXT, [metadata_HKWasUserEntered] TEXT , [device] TEXT, [metadata_HKMetadataKeyHeartRateMotionContext] TEXT, [metadata_HKDeviceManufacturerName] TEXT, [metadata_HKMetadataKeySyncVersion] TEXT, [metadata_HKMetadataKeySyncIdentifier] TEXT, [metadata_HKSwimmingStrokeStyle] TEXT, [metadata_HKVO2MaxTestType] TEXT, [metadata_HKTimeZone] TEXT, [metadata_Average HR] TEXT, [metadata_Recharge] TEXT, [metadata_Lights] TEXT, [metadata_Asleep] TEXT, [metadata_Rating] TEXT, [metadata_Energy Threshold] TEXT, [metadata_Deep Sleep] TEXT, [metadata_Nap] TEXT, [metadata_Edit Slots] TEXT, [metadata_Tags] TEXT, [metadata_Daytime HR] TEXT)

It would be nice if the columns that were added later matched the indentation of the initial columns.

I've just noticed that SQLite lets you IGNORE inserts that collide with a pre-existing key. This can be quite handy if you have a dataset that keeps changing in part, and you don't want to upsert and replace pre-existing PK rows but you do want to ignore collisions to existing PK rows.

Do sqlite_utils support such (cavalier!) behaviour?

Originally posted by @psychemedia in https://github.com/simonw/sqlite-utils/issues/18#issuecomment-480621924

Both in the Python API and the CLI tool. For the CLI tool this should work:

$ sqlite-utils create-index mydb.db mytable col1 col2

This will create a compound index across col1 and col2. The name of the index will be automatically chosen unless you use the --name=... option.

Support a --unique option too.

Right now you have to load every record into memory before passing the list to .insert_all() and friends.

If you want to process millions of rows, this is inefficient. Python has generators - we should use them!

The only catch here is that part of the magic of sqlite-utils is that it guesses the column types and creates the table for you. This code will need to be updated to notice if the table needs creating and, if it does, create it using the first X (where x=1,000 but can be customized) records.

If a record outside of those first 1,000 has a rogue column, we can crash with an error.

This will free us up to make the --nl option added in #6 much more efficient.