birdstorm
birdstorm
The problem is located that in the recent release of TiSpark, the default value of `PARTITION_PER_SPLIT` is changed from 10 to 1. It results in increasing Spark tasks. some related...
In TiDB, there are now more tables about statistics than before. e.g., some stats are not in `stats_buckets`, if the table is small, those statistics info will be in `stats_top_n`...
After #2093 and #2089, the scan performance is downgraded.
**Describe the bug**    Using longtext data type may be unable to use filter as expected. **Spark and TiSpark version info** Master
Previously, truncating decimal type for precision greater than 38 will keep the scale and truncate the precision to 38 only. It helps to keep the data's precision as much as...
**Describe the bug** Currently we are not pushing down or optimizing `in` expression. e.g., SELECT * FROM t where c1 in (1, 100, 1000) should result in a key range:...
**Describe the bug** Distinct without alias fails. **What did you do** ``` create table t(c1 int); insert into t values (2), (3); select distinct(c1), 1 as w from t; select...
**Is your feature request related to a problem? Please describe.** There isn't an elegant way to update/delete data using current API, we should investigate the new Spark 3.0 DataSourceAPIV2 for...
We could change the literal string into timestamp type, or push down casted filter so that index could be used correctly. e.g., ```sql select tp_datetime from t where tp_datetime between...
We might need to find a workaround. However, this issue is not urgent for now.