You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
in the initFetcher method of scanNode we check if we should split the given filter into 2: index-relevant and regular filters. For some special cases we do not split (i.e. don't move from out of the original filter), but rather copy index-relevant filter. This is a case for example, with array filter like _any, _all and _none. The index has no way of properly asserting these conditions, it can only give hint on what documents to check. That's why we keep the original filter intact so that it can do the final check.
With array filter within json documents the situation is the same. But it seems to be relevant only to array fields (need to double-check to be 100% sure).
For now in the method we check if fd.Kind.IsArray() || fd.Kind == client.FieldKind_NILLABLE_JSON which makes all json filters be treated this way. This is not very optimal as we can limit it only to arrays. This would mean for many json filters not on array fields, we won't need to run another filter check.
The text was updated successfully, but these errors were encountered:
in the
initFetcher
method ofscanNode
we check if we should split the given filter into 2: index-relevant and regular filters. For some special cases we do not split (i.e. don't move from out of the original filter), but rather copy index-relevant filter. This is a case for example, with array filter like_any
,_all
and_none
. The index has no way of properly asserting these conditions, it can only give hint on what documents to check. That's why we keep the original filter intact so that it can do the final check.With array filter within json documents the situation is the same. But it seems to be relevant only to array fields (need to double-check to be 100% sure).
For now in the method we check
if fd.Kind.IsArray() || fd.Kind == client.FieldKind_NILLABLE_JSON
which makes all json filters be treated this way. This is not very optimal as we can limit it only to arrays. This would mean for many json filters not on array fields, we won't need to run another filter check.The text was updated successfully, but these errors were encountered: