I’m using the
INSERT IGNORE INTO statement to insert data while ignoring duplicate pre-existing information, into a Rowstore table. Below is how the table was created:
CREATE TABLE IF NOT EXISTS rs_list (id CHAR(42), name TEXT, type1 TEXT, type2 TEXT, type3 TEXT, type4 TEXT, type5 TEXT, SHARD KEY (
type5), UNIQUE KEY (
This used to work fine in versions of MemSQL prior to 7.1, yet noticed today these tables are still showing duplicate data (where each column matches the column of another row). Is there a revised method of creating Rowstore tables to fully support the
INSERT IGNORE INTO command?
Also, has this or something similar been introduced to ColumnStores / Singlestores? I’d previously asked back in March of this year, as our operations would greatly benefit from running these outside of RowStores & freeing up tons of needed RAM.