The new options added in: - #577 Deserve consideration in the CLI as well.

https://github.com/simonw/sqlite-utils/blob/d2bcdc00c6ecc01a6e8135e775ffdb87572b802b/sqlite_utils/db.py#L1706-L1708

Found myself using this pattern to quickly prototype a schema:

```python import sqlite_utils db = sqlite_utils.Database(memory=True)

print(db["answers_chunks"].create({ "id": int, "content": str, "embedding_type_id": int, "embedding": bytes, "embedding_content_md5": str, "source": str, }, pk="id", transform=True).schema) ```

Using replace=True to drop and then recreate the table would be neat here, and would be consistent with other places that use replace=True.

Hi! Let me firstly say that I am a big fan of your work -- I follow your tweets and blog posts with great interest 😄.

Now onto the matter at hand: I think it would be great if sqlite-utils included a merge or combine command, with the purpose of combining different SQLite databases into a single SQLite database. This way, the newly "merged" database would contain all differently named tables contained in the databases to be merged as-is, as well a concatenation of all tables of the same name.

This could look something like this:

bash sqlite-utils merge cats.db dogs.db > animals.db

I imagine this is rather straightforward if all databases involved in the merge contain differently named tables (i.e. no chance of conflicts), but things get slightly more complicated if two or more of the databases to be merged contain tables with the same name. Not only do you have to "do something" with the primary key(s), but these tables could also simply have different schemas (and therefore be incompatible for concatenation to begin with).

Anyhow, I would love your thoughts on this, and, if you are open to it, work together on the design and implementation!

https://sqlite-utils.datasette.io/en/2.14.1/ isn't showing a link to the stable release right now.

I should also apply the same fix I used for Datasette in: - https://github.com/simonw/datasette/issues/1608

TIL: https://til.simonwillison.net/readthedocs/link-from-latest-to-stable

Add support for compound foreign keys, as per issue #117

Not sure if this is the right approach. In particular I'm unsure about:

• the new ForeignKey class, which replaces the namedtuple in order to ensure that column and other_column are forced into tuples. The class does the job, but doesn't feel very elegant.
• I haven't rewritten guess_foreign_table to take account of multiple columns, so it just checks for the first column in the foreign key definition. This isn't ideal.
• I haven't added any ability to the CLI to add compound foreign keys, it's only in the python API at the moment.

The PR also contains a minor related change that columns and tables are always quoted in foreign key definitions.

Hi,

I noticed the transform.table() method doesn't have an option to add/change or drop a check constraint (see https://sqlite.org/lang_createtable.html -> 3.7 Check Constraints. would be great to have this as an option!

:books: Documentation preview :books:: https://sqlite-utils--486.org.readthedocs.build/en/486/

I pushed a release for https://github.com/simonw/sqlite-utils/releases/tag/3.25.1 but forgot to include the release notes in docs/changelog.rst

This means https://sqlite-utils.datasette.io/en/stable/changelog.html isn't showing them.

Closes #398

This adds SpatiaLite helpers to the CLI.

```sh

init spatialite when creating a database

sqlite-utils create database.db --enable-wal --init-spatialite

add geometry columns

needs a database, table, geometry column name, type, with optional SRID and not-null

this will throw an error if the table doesn't already exist

sqlite-utils add-geometry-column database.db table-name geometry --srid 4326 --not-null

spatial index an existing table/column

this will throw an error it the table and column don't exist

sqlite-utils create-spatial-index database.db table-name geometry ```

Docs and tests are included.

Not sure why this has only just started failing, but I'm getting this: https://github.com/simonw/sqlite-utils/runs/4488687639

``` E sqlite3.DatabaseError: database disk image is malformed

sqlite_utils/db.py:425: DatabaseError ___ test_rebuild_fts[searchable_fts] ___

fresh_db = <Database \<sqlite3.Connection object at 0x1084ea9d0>> table_to_fix = 'searchable_fts'

@pytest.mark.parametrize("table_to_fix", ["searchable", "searchable_fts"])
def test_rebuild_fts(fresh_db, table_to_fix):
    table = fresh_db["searchable"]
    table.insert(search_records[0])
    table.enable_fts(["text", "country"])
    # Run a search
    rows = list(table.search("tanuki"))
    assert len(rows) == 1
    assert {
        "rowid": 1,
        "text": "tanuki are running tricksters",
        "country": "Japan",
        "not_searchable": "foo",
    }.items() <= rows[0].items()
    # Delete from searchable_fts_data
    fresh_db["searchable_fts_data"].delete_where()
    # This should have broken the index
    with pytest.raises(sqlite3.DatabaseError):
        list(table.search("tanuki"))
    # Running rebuild_fts() should fix it

  fresh_db[table_to_fix].rebuild_fts()

```

For https://github.com/simonw/git-history/issues/12 I found myself wanting to pass extra options to lookup() to set the column order, primary key etc.

This almost works, but it creates a rogue type record with a value of None. In [1]: import sqlite_utils In [2]: db = sqlite_utils.Database(memory=True) In [5]: db["creatures"].insert_all([ {"id": 1, "name": "Simon", "type": None}, {"id": 2, "name": "Natalie", "type": None}, {"id": 3, "name": "Cleo", "type": "dog"}], pk="id") Out[5]: <Table creatures (id, name, type)> In [7]: db["creatures"].extract("type") Out[7]: <Table creatures (id, name, type_id)> In [8]: list(db["creatures"].rows) Out[8]: [{'id': 1, 'name': 'Simon', 'type_id': None}, {'id': 2, 'name': 'Natalie', 'type_id': None}, {'id': 3, 'name': 'Cleo', 'type_id': 2}] In [9]: db["type"] Out[9]: <Table type (id, type)> In [10]: list(db["type"].rows) Out[10]: [{'id': 1, 'type': None}, {'id': 2, 'type': 'dog'}]

Originally posted by @simonw in https://github.com/simonw/sqlite-utils/issues/284#issuecomment-864416785

• Use sniff to detect CSV or TSV (if :tsv or :csv was not specified) and delimiters

Follow-on from #272

Original title: Easier way to update a row returned from .rows

Here's a surprisingly hard problem I ran into while trying to implement #239 - given a row returned by db[table].rows how can you update that row?

The problem is that the db[table].update(...) method requires a primary key. But if you have a row from the db[table].rows iterator it might not even contain the primary key - provided the table is a rowid table.

Instead, currently, you need to introspect the table and, if rowid is a primary key, explicitly include that in the select= argument to table.rows_where(...) - otherwise it will not be returned.

A utility mechanism to make this easier would be very welcome.

I came across this while about to open an issue and PR against the documentation for batch_size, which is a bit incomplete.

As mentioned in #145, while:

SQLITE_MAX_VARIABLE_NUMBER ... defaults to 999 for SQLite versions prior to 3.32.0 (2020-05-22) or 32766 for SQLite versions after 3.32.0.

it is common that it is increased at compile time. Debian-based systems, for example, seem to ship with a version of sqlite compiled with SQLITE_MAX_VARIABLE_NUMBER set to 250,000, and I believe this is the case for homebrew installations too.

In working to understand what batch_size was actually doing and why, I realized that by setting SQLITE_MAX_VARS in db.py to match the value my sqlite was compiled with (I'm on Debian), I was able to decrease the time to insert_all() my test data set (~128k records across 7 tables) from ~26.5s to ~3.5s. Given that this about .05% of my total dataset, this is time I am keen to save...

Unfortunately, it seems that sqlite3 in the python standard library doesn't expose the get_limit() C API (even though pysqlite used to), so it's hard to know what value sqlite has been compiled with (note that this could mean, I suppose, that it's less than 999, and even hardcoding SQLITE_MAX_VARS to the conservative default might not be adequate. It can also be lowered -- but not raised -- at runtime). The best I could come up with is echo "" | sqlite3 -cmd ".limits variable_number" (only available in sqlite >= 2015-05-07 (3.8.10)).

Obviously this couldn't be relied upon in sqlite_utils, but I wonder what your opinion would be about exposing SQLITE_MAX_VARS as a user-configurable parameter (with suitable "here be dragons" warnings)? I'm going to go ahead and monkey-patch it for my purposes in any event, but it seems like it might be worth considering.

Takes my test down from ten minutes to four seconds. Refs #172.

Is there a way to make .insert_all() work properly when new columns are introduced outside the first 100 records (with or without the alter=True argument)?

I'm using .insert_all() to bulk insert ~3-4k records at a time and it is common for records to need to introduce new columns. However, if new columns are introduced after the first 100 records, sqlite_utils doesn't even raise the OperationalError: table ... has no column named ... exception; it just silently drops the extra data and moves on.

It took me a while to find this little snippet in the documentation for .insert_all() (it's not mentioned under Adding columns automatically on insert/update):

The column types used in the CREATE TABLE statement are automatically derived from the types of data in that first batch of rows. Any additional or missing columns in subsequent batches will be ignored.

I tried changing the batch_size argument to the total number of records, but it seems only to effect the number of rows that are committed at a time, and has no influence on this problem.

Is there a way around this that you would suggest? It seems like it should raise an exception at least.

Got this error: File ".../python3.7/site-packages/sqlite_utils/db.py", line 462, in <dictcomp> for column in self.columns KeyError: 'REAL' .columns_dict uses REVERSE_COLUMN_TYPE_MAPPING: https://github.com/simonw/sqlite-utils/blob/43f1c6ab4e3a6b76531fb6f5447adb83d26f3971/sqlite_utils/db.py#L457-L463 REVERSE_COLUMN_TYPE_MAPPING defines FLOAT not REALA https://github.com/simonw/sqlite-utils/blob/43f1c6ab4e3a6b76531fb6f5447adb83d26f3971/sqlite_utils/db.py#L68-L74

testing some data from european power information (entsoe.eu), the title of the csv contains square brackets. as I am playing with glitch, sqlite-utils are used for creating the db.

Traceback (most recent call last):

File "/app/.local/bin/sqlite-utils", line 8, in <module>

sys.exit(cli())

File "/app/.local/lib/python3.7/site-packages/click/core.py", line 764, in call

return self.main(*args, **kwargs)

File "/app/.local/lib/python3.7/site-packages/click/core.py", line 717, in main

rv = self.invoke(ctx)

File "/app/.local/lib/python3.7/site-packages/click/core.py", line 1137, in invoke

return _process_result(sub_ctx.command.invoke(sub_ctx))

File "/app/.local/lib/python3.7/site-packages/click/core.py", line 956, in invoke

return ctx.invoke(self.callback, **ctx.params)

File "/app/.local/lib/python3.7/site-packages/click/core.py", line 555, in invoke

return callback(*args, **kwargs)

File "/app/.local/lib/python3.7/site-packages/sqlite_utils/cli.py", line 434, in insert

default=default,

File "/app/.local/lib/python3.7/site-packages/sqlite_utils/cli.py", line 384, in insert_upsert_implementation

docs, pk=pk, batch_size=batch_size, alter=alter, **extra_kwargs

File "/app/.local/lib/python3.7/site-packages/sqlite_utils/db.py", line 997, in insert_all

extracts=extracts,

File "/app/.local/lib/python3.7/site-packages/sqlite_utils/db.py", line 618, in create

extracts=extracts,

File "/app/.local/lib/python3.7/site-packages/sqlite_utils/db.py", line 310, in create_table

self.conn.execute(sql)

sqlite3.OperationalError: unrecognized token: "]"

entsoe_2016.csv

renamed to txt for uploading compatibility

entsoe_2016.txt

code is remixed directly from your https://glitch.com/edit/#!/datasette-csvs repo

The foreign_keys= argument to table.insert_all() and friends can be used to specify foreign key relationships that should be created.

It is not yet documented. It also requires you to specify the SQLite type of each column, even though this can be detected by introspecting the referenced table:

cols = [c for c in self.db[other_table].columns if c.name == other_column]
cols[0].type

Relates to #2