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
Without this, we will either need to modify how the JS UI handles CSV download of very large tables, or will need to find another way to slice up content, like with head/tail calls before requesting a snapshot.
DoGet does not support extra metadata, so we can't just attach a custom barrage object with an index for rows and a bitset for columns.
The text was updated successfully, but these errors were encountered:
Head and Tail is probably not enough. The selected rows do not need to be contiguous; if you break this up into one DoGet per contiguous range then you still can't get a consistent snapshot on a ticking table.
Also, one parameter I want to add to the "DoGet over DoExchange" is the ability to control chunk size. Since DoGet/DoExchange are modeled as a stream of record batches, it becomes very easy to respect a user requested batch size.
…#1973)
This uses the flight DoExchange call with the BarrageSnapshotRequest
payload.
Also fixes leak of rowsets in BarrageStreamGenerator, and correctly
reports the RecordBatch length for viewport snapshots.
Fixes#1678Fixes#1039
Without this, we will either need to modify how the JS UI handles CSV download of very large tables, or will need to find another way to slice up content, like with head/tail calls before requesting a snapshot.
DoGet does not support extra metadata, so we can't just attach a custom barrage object with an index for rows and a bitset for columns.
The text was updated successfully, but these errors were encountered: