Add support for `_varchar`

[universalmind303]

Describe your feature request

if you have a postgres table with a varchar[] type, connectorx will panic

thread '<unnamed>' panicked at 'not implemented: _varchar',

It would be nice if connectorx supported _varchar types.

[Jordan-M-Young]

Hey @universalmind303 I've opened a PR for this issue here

[davidjbrennan]

Would be great if connectorx also supported _text (i.e. postgres tables using columns of type text[]). Connectorx suffers from the same "not implemented" as _varchar.

[wangxiaoying]

Hi @universalmind303 , @davidjbrennan , our latest alpha release support _varchar to arrow2 thanks to the contribution from @Jordan-M-Young #487.

You can install via pip install connectorx==0.3.2a5 and have a try. This feature is not directly supported in pandas destination yet. If you need pandas as output, one workaround is to get the arrow2 destination first and then convert arrow to pandas.

[davidjbrennan]

Hi @Jordan-M-Young. It would be great if you could extend the work you did to support _varchar to arrow2 to also include _text to arrow2.

[Jordan-M-Young]

Hey @davidjbrennan,

I think I may have some time to do that work coming up. Do you have an issue open by chance?

[davidjbrennan]

Hi @Jordan-M-Young. No I don't have an issue open as I thought that resolving the _text issue might be just a slight expansion (TEXT is like declaring VARCHAR without a length specifier) of this currently open _varchar issue.

[Jordan-M-Young]

@davidjbrennan All good. Yea you're right it should be relatively straightforward. I'll take a crack at it.

[Jordan-M-Young]

@davidjbrennan I've opened a PR for _text support.

[davidjbrennan]

I really appreciate you spending time on this @Jordan-M-Young. I'm looking forward when I can download and use it.

[Jordan-M-Young]

Hi @davidjbrennan looks like the _text PR has been merged.