Issue with merging explore results due to row limits

[Federico1]

Hello,

I was trying out the ‘Merge Results’ functionality in an Explore and I couldn’t make it work properly for large tables.

It seems that the row limit (5k maximum) applied in the individual explore results are affecting the ability to merge them together.

As an example: suppose I want to merge an explore with two dimensions (uuid, name) with an identical copy of itself (still uuid, name) on uuid=uuid.
This works perfectly as long as I order the two explores the same way (by clicking on the uuid header).
If I invert the ordering between the two columns (the first explore is ascending, the second is descending) suddenly my merge result will contain all nulls for the second explore.

As mentioned above, my hypothesis is that the row limit is applied individually to both explores before the merge meaning that, unless ordered the same way, the second explore may not have the necessary uuids in its first [limit] rows.

Could someone help me confirm whether this is a bug or if its actually the intended behaviour?

Best,
Federico

edit: typos

[izzymiller]

This is the intended behavior, which means you have to be super careful when making merges that contain more than 1k rows. Our docs give it a call out near the beginning:

Looker Documentation

Merging Results from Different Explores

Combine data from multiple Explores using Merged Results.

You’re spot on that ordering by the same field is the best way to do this— But bear in mind that even if you order on the right field, if you have more than 5,000 rows you may not get all of the matches that actually exist. I try to always group explores such that they’re <5k before I merge.