[Doris On Elasticsearch] Develop Roadmap

Community users are very welcome to actively participate in the development!

• Move the Predicate-Processing from Doris BE to Doris FE @wuyunfeng

• directly process the CastExpr @wuyunfeng

• support function push-down，such as a = b

• add QueryStringQuery for analyzed field, gradually discard es_query

• improve the performance under muti-table joins, provide more data/index statistic infomation

• push down aggreate operator to Elasticsearch such as count/avg/sum etc under single index context

• add the ability for process object and nested type field

• fetch index mapping and shard location divided into two/three phase @stalary

  1. GET {index}/{type}/_mapping

  2. GET _nodes/http

  3. GET {index}/_search_shards

• add ability for processing wildcard and alias index @wuyunfeng

• support alter operation for ES table @HappenLee

• Extended sql such as load all table from Elasticsearch, get rid of creating table manually

• Add Debug mode for ES table @wuyunfeng

• When reading more column values such as above 20, switch from column reading to row reading @wuyunfeng

• Support ES partition to meet the demand of auto-deleting some data out of data in ES @blackfox1983

[stalary]

index mapping和shard location的检测分成两个阶段 这个我来搞下

[wuyunfeng]

https://github.com/apache/incubator-doris/issues/3249