I’m trying to “debug” an issue where LEFT join seems to be taking unprecedentedly long to complete. Following on from Crate’s documentation on how to write joins (and the usual SQL) I wanted to join two tables and find out which records I was missing from the other table.
SELECT * FROM TABLE1 T1
LEFT JOIN TABLE T2
ON T1.COLUMN1 = T2.COLUMN1
WHERE T2.COLUMN1 IS NULL
I had to fall back on the traditional SQL (MSSQL) and have run this query on the same dataset with no problems what so ever. I know Crate handles joins differently than traditional relational database would, I was wondering if anyone has faced this issue and there was something I could do or write the query in a different way? (I tried using IN approach but to no avail)
(asked on SO https://stackoverflow.com/questions/56209606/cratedb-left-join-performance)