[BUG] tispark 3.x版本读取tikv时,读取并发比2.5版本少,速度也慢

这个执行计划应该是有问题的,明显没有用到索引,怎么会走索引呢。这是tidb的执行计划:

|Projection_4|3462592.98|root||qk_chi.perio_art_project.project_name, Column#11|
|---|---|---|---|---|
|└─HashAgg_9|3462592.98|root||group by:qk_chi.perio_art_project.project_name, funcs:count(Column#12)->Column#11, funcs:firstrow(qk_chi.perio_art_project.project_name)->qk_chi.perio_art_project.project_name|
|  └─TableReader_10|3462592.98|root||data:HashAgg_5|
|    └─HashAgg_5|3462592.98|cop[tikv]||group by:qk_chi.perio_art_project.project_name, funcs:count(1)->Column#12|
|      └─TableFullScan_8|18072492.00|cop[tikv]|table:perio_art_project|keep order:false|