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
A discussion started from this post in #1497 raised a question: should we implement this feature or remove it (revert)?
If we keep this feature
Procs
Compatibility with legacy engine
Proper use of total and size in the response - these values were always equal before
Cons
Rework required, implementation is not good
It is not used; main consumer of JDBC response format is the JDBC driver which ignores both these values
If we remove this feature
Procs
Reduce impact of pagination feature
Cons
A breaking change: legacy engine properly sets total and size for paging requests
Note: it is still not used by JDBC (ref: JsonQueryResponse).
Please, feel free to edit procs/cons lists above and post your votes. Removal was already discussed in meetings and in the slack channel, but should be posted/copied here for visibility.
The text was updated successfully, but these errors were encountered:
if the request from JDBC, execute in new engine, ignore total field.
What does mean ignore? Return total equal to size as it was before (breaking with V1) or not to return it at all (breaking at all)?
if the request from others, fallback to legacy.
That means any paging request in non-jdbc format will be processed in V1. Why?
Note: legacy engine does not support paging for non-jdbc format too - it returns non-paged results (aka ignores fetch_size).
Is your feature request related to a problem?
A discussion started from this post in #1497 raised a question: should we implement this feature or remove it (revert)?
If we keep this feature
Procs
total
andsize
in the response - these values were always equal beforeCons
If we remove this feature
Procs
Cons
total
andsize
for paging requestsNote: it is still not used by JDBC (ref: JsonQueryResponse).
Please, feel free to edit procs/cons lists above and post your votes. Removal was already discussed in meetings and in the slack channel, but should be posted/copied here for visibility.
The text was updated successfully, but these errors were encountered: