table.update(...) method

[simonw]

Spun off from #23 - this method will allow a user to update a specific row.

Currently the only way to do that it is to call .upsert({full record}) with the primary key field matching an existing record - but this does not support partial updates.

db["events"].update(3, {"name": "Renamed"})

This method only works on an existing table, so there's no need for a pk="id" specifier - it can detect the primary key by looking at the table.

If the primary key is compound the first argument can be a tuple:

db["events_venues"].update((3, 2), {"custom_label": "Label"})

The method can be called without the second dictionary argument. Doing this selects the row specified by the primary key (throwing an error if it does not exist) and remembers it so that chained operations can be carried out - see proposal in #23 (comment)

[simonw]

The last touched primary key should go in last_pk.

sqlite-utils/docs/python-api.rst

Line 126 in ba6bd47

The ``dogs.last_pk`` property will return the last inserted primary key value, if you specified one. This can be very useful when writing code that creates foreign keys or many-to-many relationships.

[simonw]

Should this support both dictionaries AND keyword arguments?

db["events_venues"].update((3, 2), custom_label="Label")

Pros: this is quite nice to look at.
Cons: would have to continue supporting dictionaries as well, since otherwise you won't be able to update columns with names that are not valid Python identifiers. So now there are two ways.

I'm going to stick with dictionaries only for the moment.