The REFRESH statement reloads the metadata for the table from the metastore database and does an incremental reload of the file and block metadata from the HDFS NameNode. REFRESH is used to avoid inconsistencies between Impala and external metadata sources, namely Hive Metastore (HMS) and NameNodes.
The REFRESH statement is only required if you load data from outside of Impala. Updated metadata, as a result of running REFRESH, is broadcast to all Impala coordinators.
See Overview of Impala Metadata and the Metastore for the information about the way Impala uses metadata and how it shares the same metastore database as Hive.
REFRESH [db_name.]table_name [PARTITION (key_col1=val1 [, key_col2=val2...])]
The table name is a required parameter, and the table must already exist and be known to Impala.
Only the metadata for the specified table is reloaded.
Use the REFRESH statement to load the latest metastore metadata for a particular table after one of the following scenarios happens outside of Impala:
- Deleting or adding files.
For example, after loading new data files into the HDFS data directory for the table, including inserting data from Hive via INSERT or LOAD DATA.
- Deleting, adding, or modifying partitions.
For example, after issuing ALTER TABLE or other table-modifying SQL statement in Hive
- INVALIDATE METADATA is an asynchronous operations that simply discards the loaded metadata from the catalog and coordinator caches. After that operation, the catalog and all the Impala coordinators only know about the existence of databases and tables and nothing more. Metadata loading for tables is triggered by any subsequent queries.
- REFRESH reloads the metadata synchronously. REFRESH is more lightweight than doing a full metadata load after a table has been invalidated. REFRESH cannot detect changes in block locations triggered by operations like HDFS balancer, hence causing remote reads during query execution with negative performance implications.
Refreshing a single partition:
In CDH 5.9 / Impala 2.7 and higher, the REFRESH statement can apply to a single partition at a time, rather than the whole table. Include the optional PARTITION (partition_spec) clause and specify values for each of the partition key columns.
- The PARTITION clause of the REFRESH statement must include all the partition key columns.
- The order of the partition key columns does not have to match the column order in the table.
- Specifying a nonexistent partition does not cause an error.
- The partition can be one that Impala created and is already aware of, or a new partition created through Hive.
The following examples demonstrates the above rules.
-- Partition doesn't exist. REFRESH p2 PARTITION (y=0, z=3); REFRESH p2 PARTITION (y=0, z=-1) -- Key columns specified in a different order than the table definition. REFRESH p2 PARTITION (z=1, y=0) -- Incomplete partition spec causes an error. REFRESH p2 PARTITION (y=0) ERROR: AnalysisException: Items in partition spec must exactly match the partition columns in the table definition: default.p2 (1 vs 2)
For examples of using REFRESH and INVALIDATE METADATA with a combination of Impala and Hive operations, see Switching Back and Forth Between Impala and Hive.
Related impala-shell options:
Due to the expense of reloading the metadata for all tables, the impala-shell -r option is not recommended.
All HDFS and Sentry permissions and privilege requirements are the same whether you refresh the entire table or a single partition.
The REFRESH statement checks HDFS permissions of the underlying data files and directories, caching this information so that a statement can be cancelled immediately if for example the impala user does not have permission to write to the data directory for the table. Impala reports any lack of write permissions as an INFO message in the log file.
If you change HDFS permissions to make data readable or writeable by the Impala user, issue another REFRESH to make Impala aware of the change.
Much of the metadata for Kudu tables is handled by the underlying storage layer. Kudu tables have less reliance on the metastore database, and require less metadata caching on the Impala side. For example, information about partitions in Kudu tables is managed by Kudu, and Impala does not cache any block locality metadata for Kudu tables.
The REFRESH and INVALIDATE METADATA statements are needed less frequently for Kudu tables than for HDFS-backed tables. Neither statement is needed when data is added to, removed, or updated in a Kudu table, even if the changes are made directly to Kudu through a client program using the Kudu API. Run REFRESH table_name or INVALIDATE METADATA table_name for a Kudu table only after making a change to the Kudu table schema, such as adding or dropping a column.