Erez Shinan

You can do `--json > /home/user/dir/output.json`

This feature hasn't been implemented yet, but it's definitely something we are going to support.

@yuzeh Here is a good start :)

@avaitla @vlad-bezden @yuzeh The PR for a compound primary key is ready! You can install it with: ```shell pip install git+https://github.com/datafold/data-diff.git@refs/pull/158/head ``` And then you can use multiple `-k` arguments,...

@vlad-bezden @yuzeh @avaitla I had some time to think about this problem, and I think it should work with the current version (not the new PR), by setting `-k` to...

@yuzeh Can you please share the SQL type of the column you use as a key, and a few sample values from that column?

# Composite Keys - Analysis of problems, and proposed solutions Data-diff's performance relies strongly on its ability to split the table rows into (more-or-less) evenly sized segments. I see two...

Just a note - it may be possible to infer the schema from looking at the data, at least in some cases. We currently do this to detect UUIDs, which...

Looks like there's a version collision on numpy, with one of the other optional drivers. Is there a way to avoid this nested requirement? Also, once it's stable, please rebase...

@nklsw The tests fail unless I import the csv, since the schema needs to be created first. But also, when using the test connection string, calling "create schema" fails with...