Is it still nfortunately slow and tricky when playing with floats ?

Not sure what you mean here?

Testing bulk --functions:

% sqlite-utils create-table chickens.db chickens id integer name text name_upper text % echo 'id,name 1,Blue 2,Snowy 3,Azi 4,Lila 5,Suna 6,Cardi' | sqlite-utils bulk chickens.db ' insert into chickens (id, name, name_upper) values (:id, :name, myupper(:name)) ' - --functions ' def myupper(s): return s.upper() ' --csv % sqlite-utils rows chickens.db chickens [{"id": 1, "name": "Blue", "name_upper": "BLUE"}, {"id": 2, "name": "Snowy", "name_upper": "SNOWY"}, {"id": 3, "name": "Azi", "name_upper": "AZI"}, {"id": 4, "name": "Lila", "name_upper": "LILA"}, {"id": 5, "name": "Suna", "name_upper": "SUNA"}, {"id": 6, "name": "Cardi", "name_upper": "CARDI"}]

I don't think I need separate documentation for sqlite-utils memory and sqlite-tils bulk since they work the same, and the --help text provides the necessary hints.

Documentation: https://sqlite-utils.datasette.io/en/latest/cli.html#defining-custom-sql-functions

I should add this --functions feature to the memory and bulk commands too.

Found a fix for that!

I replaced this:

python locals = {} globals = {} exec(functions, globals, locals) # Register all callables in the locals dict: for name, value in locals.items(): if callable(value): db.register_function(value, name=name) With this: python globals = {} exec(functions, globals) # Register all callables in the globals dict: for name, value in globals.items(): if callable(value): db.register_function(value, name=name) Because https://docs.python.org/3/library/functions.html#exec says:

If only globals is provided, it must be a dictionary (and not a subclass of dictionary), which will be used for both the global and the local variables. If globals and locals are given, they are used for the global and local variables, respectively.

This works: ``` sqlite-utils :memory: 'select extract_domain("https://www.google.com/blah")' --functions ' from urllib.parse import urlparse

def extract_domain(url): global urlparse return urlparse(url).netloc ' ```

I've encountered this problem before: https://sqlite-utils.datasette.io/en/stable/cli.html#cli-convert-complex

``` $ sqlite-utils convert content.db articles score ' import random random.seed(10)

def convert(value): global random return random.random() ' `` Note theglobal randomline here. Due to the way the tool compiles Python code, this is necessary to ensure therandommodule is available within theconvert()function. If you were to omit this you would see aNameError: name 'random' is not defined` error.

I tried this: ``` sqlite-utils :memory: 'select extract_domain("https://www.google.com/blah")' --functions ' from urllib.parse import urlparse

def extract_domain(url): from urllib.parse import urlparse return urlparse(url).netloc ' And got: NameError: name 'urlparse' is not defined Error: user-defined function raised exception But this worked OK: % sqlite-utils :memory: 'select extract_domain("https://www.google.com/blah")' --functions ' def extract_domain(url): from urllib.parse import urlparse return urlparse(url).netloc ' [{"extract_domain(\"https://www.google.com/blah\")": "www.google.com"}] ```

Quick prototype: ```diff diff --git a/sqlite_utils/cli.py b/sqlite_utils/cli.py index 43e76fa..5dee4f6 100644 --- a/sqlite_utils/cli.py +++ b/sqlite_utils/cli.py @@ -1633,6 +1633,9 @@ def drop_view(path, view, ignore, load_extension): type=(str, str), help="Named :parameters for SQL query", ) +@click.option( + "--functions", help="Python code defining one or more custom SQL functions" +) @load_extension_option def query( path, @@ -1649,6 +1652,7 @@ def query( raw, param, load_extension, + functions, ): """Execute SQL query and return the results as JSON

@@ -1665,6 +1669,16 @@ def query( _load_extensions(db, load_extension) db.register_fts4_bm25()

• Register any Python functions as SQL functions:

• if functions:
• locals = {}
• globals = {}
• exec(functions, globals, locals)

• Register all callables in the locals dict:

• for name, value in locals.items():
• if callable(value):
• db.register_function(value, name=name) + _execute_query( db, sql, param, raw, table, csv, tsv, no_headers, fmt, nl, arrays, json_cols ) Demo:bash % sqlite-utils :memory: 'select 1 + dog()' --functions ' quote> def dog(): quote> return 2 quote> ' [{"1 + dog()": 3}] ```

Alternative design would be --function name definition - like this:

sqlite-utils data.db 'update images set domain = extract_domain(url)' --function extract_domain ' from urllib.parse import urlparse return urlparse(url).netloc ' I like the --functions design better because it leaves space for import statements at the top of the code block, and allows more than one function to be defined in a single go.