Snappydata SQL WITH statement -
i using snappydata run queries, , use sql statement:
with x ( select distinct col_a, col_b table_a ) insert table_b select x.col_a, x.col_b x join table_c c on x.col_a = c.col_a , x.col_b = c.col_b
this sql code runs fine when running in local mode, when submit compiled jar file snappydata cluster, throws error saying table "app.x" not exist,
org.apache.spark.sql.tablenotfoundexception: table 'app.x' not found;
any idea why happens?
this known limitation due approximate query processing (aqp) rule resolution tries resolve name prematurely. saw when running tpc-ds queries. fixed on latest aqp master , fix available in 1.0. there 2 options:
a) change normal sub-query variant:
insert table_b select x.col_a, x.col_b (select distinct col_a, col_b table_a) x join table_c c on x.col_a = c.col_a , x.col_b = c.col_b
b) or if not using aqp queries, can move out snappydata-aqp jar product jars directory
Comments
Post a Comment