SparkSQL doesn't support offset. There's a hack based on Sequence number annotation that
we used to use in Mimir 1, but that's no longer present/needed in Mimir 2.
https://github.com/UBOdin/mimir-api/commit/bfb7236af6dd26ffc1daf2bd747379c79b809619 introduced the
/query/table endpoint in the Mimir API, which allows direct lookups to a table / view definition,
including paged access via offset/limit. This push migrates the main Mimir datset accessors
(i.e., vizier.datastore.mimir.dataset) to this new table API. As a side effect, this should
make the table lookups more resilient to naming issues in tables/columns, since we no longer have
to go through SQL and the associated challenges in escaping/capitalizing identifiers.
(closes https://github.com/VizierDB/web-ui/issues/221)
SparkSQL doesn't support offset. There's a hack based on Sequence number annotation that we used to use in Mimir 1, but that's no longer present/needed in Mimir 2.
https://github.com/UBOdin/mimir-api/commit/bfb7236af6dd26ffc1daf2bd747379c79b809619 introduced the /query/table endpoint in the Mimir API, which allows direct lookups to a table / view definition, including paged access via offset/limit. This push migrates the main Mimir datset accessors (i.e., vizier.datastore.mimir.dataset) to this new table API. As a side effect, this should make the table lookups more resilient to naming issues in tables/columns, since we no longer have to go through SQL and the associated challenges in escaping/capitalizing identifiers.