Hi, is there any reported/known issue with the export documents mechanism? It seems not all properties of the records are being extracted! I'm using the CURL command like this

curl -H "X-TYPESENSE-API-KEY: ${DEV_API_KEY}" "${DEV_BASE_URL}/collections/${DEV_COLLECTION_NAME}/documents/export" -o "${TS_EXPORTED_FILE}"

No, there are no known issues. Can you provide a simple reproducible example that we can try with?

I'm not sure how can I share any reproducible example as I can't share clients data. In the end I ended up with collecting all the field's name mentioned in the schema and append them with the CURL request. A note that might be relevant - we a facet field defined with wild card pattern so that we can have facets generated dynamically based on data

If you can create a collection with a single document (you can replace actual data with dummy values) and then show what's missing from export, we can debug.

We had a similar case before – embeddings aren’t exported by default, so we had to explicitly include the vector fields for export as well.

Oh I see, I will check.

Thanks, but it looks like a feature since embeddings are usually huge, so it makes sense to export them on demand :)

Right, yeah we exclude that normal search and doc fetch endpoint and i think that's carrying over to export as well.

Yeah, same issue here. It didn't happen in the previous version

We've excluded embedding fields in previous version also

In 0.26 I was fetching it with the export, without adding nothing

For export or for search APIs?

export

Ideally there should be no limit, I will check and get back to you.

Thank you!

We've fixed the timeout issue with export + filter_by in a recent commit. Will also fix the embeddings fields being excluded since that's a surprising, unintended behavior.

Nice! Which rc?

The timeout issue with export and as well as missing fields is fixed in

28.0.rc5