- Sep 16, 2024
-
-
EmirVildanov authored
-
EmirVildanov authored
-
EmirVildanov authored
-
EmirVildanov authored
-
EmirVildanov authored
-
EmirVildanov authored
-
EmirVildanov authored
-
Denis Smirnov authored
-
Denis Smirnov authored
-
Denis Smirnov authored
Co-authored-by:
Georgy Moshkin <gmoshkin@picodata.io>
-
- Sep 13, 2024
-
-
Previously some public lua functions in core-router.lua failed returning nil instead of an error. As a result in a case of error rust code treated it as success, but failed to pop expected results from lua stack. So, instead of a real error user got lua reading errors. Fixed. Co-authored-by:
Denis Smirnov <sd@picodata.io>
-
godzie44 authored
-
EmirVildanov authored
-
EmirVildanov authored
-
EmirVildanov authored
-
EmirVildanov authored
-
EmirVildanov authored
-
EmirVildanov authored
-
EmirVildanov authored
feat: add logic of correct local SQL generation for the case of asterisk (not in a view of references, but as "*")
-
EmirVildanov authored
-
EmirVildanov authored
-
EmirVildanov authored
-
EmirVildanov authored
-
EmirVildanov authored
fix: remove useless SubQuery nodes, fix logic of vtable columns naming (automatic indexes instead of names from local SQL metadata)
-
- Sep 12, 2024
-
-
Maksim Kaitmazian authored
-
Andrey Strochuk authored
-
- Sep 11, 2024
-
-
Andrey Strochuk authored
-
Andrey Strochuk authored
-
Andrey Strochuk authored
-
- Sep 10, 2024
-
-
-
Kurdakov Alexander authored
-
- Sep 09, 2024
-
-
- For cacheable requests, we now send only cache required data needed to execute the query if it is in the storage cache. If query is not in the storage cache, storage returns empty result with flag indicating there was a cache miss. Then coordinator sends the second round of requests to replicasets where cache miss happened. It sends the full message there. - Parameters for required message are now computed without compilation to SyntaxPlan - Fixed parameter order in SyntaxPlan
-
- Sep 06, 2024
-
-
-
Denis Smirnov authored
-
Denis Smirnov authored
-
Denis Smirnov authored
-
Denis Smirnov authored
-
Denis Smirnov authored
-
Denis Smirnov authored
-
Denis Smirnov authored
Previously we cached only the leaf sub-trees of the IR plan under the first level of motions. The problem was in temporary tables as we used random names for them and dropped after usage (and invalidated corresponding prepared statements as well). Now we use idempotent names for temporary tables and truncate them after usage. It allows us to keep prepared statements in the cache. One more thing: as we use the same temporary table for multiple queries, we need to use latch locks for the prepared statement cache. One day we'll switch to more granular locks (mostly for vinyl spaces), but let's use current approach for the start. We also removed can_be_cached flag from the code. Only DQL uses VDBE and needs caching, and all these queries can be cached. The only exception are the queries with Local and LocalSegment motion policies but only DML can have them. So, no need for this flag and let's make the code cleaner.
-