MongoDB is schema-less, so field may have different types. When dealing with such databases, the $sample stage may produce different results with each execution.
I think it's bothersome to have differences with each execution, so it would be preferable to either suppress this or output all types of the aggregated documents.
What this PR does
From the documents selected in the $sample stage, all types of the fields will be enumerated with commas separated.
This can be enabled with the multipleFieldType query.
As one solution, I've implemented this feature, but holding the types in a comma-separated format might be something we need to discuss. I would like to hear your opinion.
Issue Context / why we need it
MongoDB is schema-less, so field may have different types. When dealing with such databases, the
$sample
stage may produce different results with each execution.For example:
I think it's bothersome to have differences with each execution, so it would be preferable to either suppress this or output all types of the aggregated documents.
What this PR does
From the documents selected in the
$sample
stage, all types of the fields will be enumerated with commas separated.This can be enabled with the
multipleFieldType
query.For example:
As one solution, I've implemented this feature, but holding the types in a comma-separated format might be something we need to discuss. I would like to hear your opinion.