issue_comments

Now live at https://sqlite-utils.datasette.io/en/latest/installation.html#setting-up-shell-completion

Now live at https://sqlite-utils.datasette.io/en/latest/cli.html

Codecov Report

Patch and project coverage have no change.

Comparison is base (e240133) 96.36% compared to head (0b81794) 96.36%.

:exclamation: Current head 0b81794 differs from pull request most recent head 21036a5. Consider uploading reports for the commit 21036a5 to get more accurate results

:umbrella: View full report in Codecov by Sentry.
:loudspeaker: Do you have feedback about the report comment? Let us know in this issue.

If https://sqlite-utils--553.org.readthedocs.build/en/553/cli.html#running-sql-queries looks good I can merge this.

This page has all of the changes: https://sqlite-utils--553.org.readthedocs.build/en/553/cli.html#running-sql-queries

Put this in a PR so I can preview it: - #553

Now released: https://sqlite-utils.datasette.io/en/stable/changelog.html#v3-32

Another option:

That's using this markup:

``` Newline-delimited JSON ~~~~~~~~~~~~~~~~~~~~~~

Use --nl to get back newline-delimited JSON objects:

.. code-block:: bash

sqlite-utils dogs.db "select * from dogs" --nl

.. code-block:: output

{"id": 1, "age": 4, "name": "Cleo"}
{"id": 2, "age": 2, "name": "Pancakes"}

``` And this extra CSS:

css .highlight-output .highlight { border-left: 9px solid #30c94f; }

I could split them up into two blocks like this:

I do miss the visual indication that one of these is the command and one is the output though.

I wrote the docs like this because early examples include both the command and its output:

https://sqlite-utils.datasette.io/en/stable/cli.html#returning-json

Tests passed here: https://github.com/simonw/sqlite-utils/actions/runs/5039119716

Codecov Report

Patch coverage: 83.33% and project coverage change: +0.06 :tada:

Comparison is base (b3b100d) 96.30% compared to head (948692a) 96.36%.

:umbrella: View full report in Codecov by Sentry.
:loudspeaker: Do you have feedback about the report comment? Let us know in this issue.

Relevant issues: - https://github.com/python/importlib_metadata/issues/406 - https://github.com/PyCQA/flake8/issues/1701

It looks to me like this is only a problem for flake8 on Python 3.7 - 3.8 and higher work OK.

This is blocking: - #549

Draft documentation: https://sqlite-utils--549.org.readthedocs.build/en/549/cli.html#cli-tui

Failing mypy test: https://github.com/simonw/sqlite-utils/actions/runs/5038983349/jobs/9036828465 sqlite_utils/cli.py:37: error: Skipping analyzing "trogon": module is installed, but missing library stubs or py.typed marker [import] sqlite_utils/cli.py:37: note: See https://mypy.readthedocs.io/en/stable/running_mypy.html#missing-imports Found 1 error in 1 file (checked 52 source files)

Documentation: https://sqlite-utils--549.org.readthedocs.build/en/549/cli.html#cli-tui

Oh, I see why that is now:

https://github.com/simonw/sqlite-utils/blob/6027f3ea6939a399aeef2578fca17efec0e539df/sqlite_utils/cli.py#L2670-L2679

This is because of the following command:

sqlite-utils analyze-tables table1 table2 --column x

Since you can pass multiple tables AND multiple columns, the tool currently assumes that the column(s) you specify may be available on a subset of the provided tables.

I'm going to change this so if the column is not on ANY of those tables you get an error.

This will be a cosmetic change to the CLI output only - the options to save data to the database and the Python API function will continue to return [(None, 158)].

I can add an optimization though to avoid running the SQL count query if we know that it's all null.

New docs:

• https://sqlite-utils.datasette.io/en/latest/cli.html#cli-analyze-tables
• https://sqlite-utils.datasette.io/en/latest/cli-reference.html#analyze-tables
• https://sqlite-utils.datasette.io/en/latest/python-api.html#analyzing-a-column
• https://sqlite-utils.datasette.io/en/latest/reference.html#sqlite_utils.db.Table.analyze_column

New help output:

``` % sqlite-utils analyze-tables --help Usage: sqlite-utils analyze-tables [OPTIONS] PATH [TABLES]...

Analyze the columns in one or more tables

Example:

  sqlite-utils analyze-tables data.db trees

Options: -c, --column TEXT Specific columns to analyze --save Save results to _analyze_tables table --common-limit INTEGER How many common values --no-most Skip most common values --no-least Skip least common values --load-extension TEXT Path to SQLite extension, with optional :entrypoint -h, --help Show this message and exit. ```

Documentation preview: - https://sqlite-utils--546.org.readthedocs.build/en/546/cli.html#cli-analyze-tables - https://sqlite-utils--546.org.readthedocs.build/en/546/cli-reference.html#analyze-tables - https://sqlite-utils--546.org.readthedocs.build/en/546/python-api.html#analyzing-a-column - https://sqlite-utils--546.org.readthedocs.build/en/546/reference.html#sqlite_utils.db.Table.analyze_column

Codecov Report

Patch coverage: 93.75% and no project coverage change.

Comparison is base (b3b100d) 96.30% compared to head (9f23e68) 96.31%.

:exclamation: Current head 9f23e68 differs from pull request most recent head 2eca17d. Consider uploading reports for the commit 2eca17d to get more accurate results

:umbrella: View full report in Codecov by Sentry.
:loudspeaker: Do you have feedback about the report comment? Let us know in this issue.

I generated the commit message in https://github.com/simonw/sqlite-utils/commit/1c1991b447a1ddd3d61d9d4a8a1d6a9da47ced20 using git diff | llm --system 'describe this change'.

It looks like nargs=-1 on a positional argument isn't yet supported - opened an issue here: - https://github.com/Textualize/trogon/issues/4

Opened a feature request for customizing the help and command name:

• https://github.com/Textualize/trogon/issues/2

OK, this works!

To try it out, install that branch from GitHub:

pip install https://github.com/simonw/sqlite-utils/archive/refs/heads/trogon.zip

Then run this:

sqlite-utils install trogon

And this:

sqlite-utils tui

I don't want to add trogon as a default dependency because it's a little heavy - it pulls in all of Rich and Textual as well. People who use sqlite-utils just for its Python API won't benefit from this - it's a CLI feature only.

But I have a sqlite-utils install ... command for helping people to install packages into the same virtual environment as sqlite-utils no matter how they installed that tool: https://sqlite-utils.datasette.io/en/stable/cli.html#cli-install

So I can treat Trogon as an optional dependency and add the sqlite-utils tui command only if that package is also installed.