UPSERT Statement (Impala 2.8 or higher only)

Acts as a combination of the INSERT and UPDATE statements. For each row processed by the UPSERT statement:

  • If another row already exists with the same set of primary key values, the other columns are updated to match the values from the row being “UPSERTed”.

  • If there is not any row with the same set of primary key values, the row is created, the same as if the INSERT statement was used.

This statement only works for Impala tables that use the Kudu storage engine.

Syntax:

  2. UPSERT [hint_clause] INTO [TABLE] [db_name.]table_name
  3.   [(column_list)]
  4. {
  5.     [hint_clause] select_statement
  6.   | VALUES (value [, value ...]) [, (value [, value ...]) ...]
  7. }
  9. hint_clause ::= [SHUFFLE] | [NOSHUFFLE]
  10.   (Note: the square brackets are part of the syntax.)

The select_statement clause can use the full syntax, such as WHERE and join clauses, as SELECT Statement.

Statement type: DML

Usage notes:

If you specify a column list, any omitted columns in the inserted or updated rows are set to their default value (if the column has one) or NULL (if the column does not have a default value). Therefore, if a column is not nullable and has no default value, it must be included in the column list for any UPSERT statement. Because all primary key columns meet these conditions, all the primary key columns must be specified in every UPSERT statement.

Because Kudu tables can efficiently handle small incremental changes, the VALUES clause is more practical to use with Kudu tables than with HDFS-based tables.

Important: After adding or replacing data in a table used in performance-critical queries, issue a COMPUTE STATS statement to make sure all statistics are up-to-date. Consider updating statistics for a table after any INSERT, LOAD DATA, or CREATE TABLE AS SELECT statement in Impala, or after loading data through Hive and doing a REFRESH table_name in Impala. This technique is especially important for tables that are very large, used in join queries, or both.

Examples:

  2. UPSERT INTO kudu_table (pk, c1, c2, c3) VALUES (0, 'hello', 50, true), (1, 'world', -1, false);
  3. UPSERT INTO production_table SELECT * FROM staging_table;
  4. UPSERT INTO production_table SELECT * FROM staging_table WHERE c1 IS NOT NULL AND c2 > 0;

Related information:

Using Impala to Query Kudu Tables, INSERT Statement, UPDATE Statement (Impala 2.8 or higher only)

Parent topic: Impala SQL Statements