merge_chunks()

Merge two or more chunks into one chunk

Since 2.18.0

Merge two or more s into one.

The partition boundaries for the new is the union of all partitions of the merged s. The new retains the name, constraints, and triggers of the first in the partition order.

You can only merge s that have directly adjacent partitions. It is not possible to merge s that have another , or an empty range between them in any of the partitioning dimensions.

merging has the following limitations. You cannot:

• Merge s with tiered data
• Write to s that are being merged

Concurrent mode

When a merge is executed using the concurrently option, other processes can simultaneously read from the s being merged and insert into other s. The merge happens across two transactions: the first one rewrites the s into a temporary relation without taking any locks that prevent reads, while the second transaction locks out all other operations before swapping the old relations for the new one. The second operation completes quickly so it should not significantly affect other operations.

Samples

• Merge two s:

  CALL merge_chunks('_timescaledb_internal._hyper_1_1_chunk', '_timescaledb_internal._hyper_1_2_chunk');
  

• Merge more than two s:

  CALL merge_chunks('{_timescaledb_internal._hyper_1_1_chunk, _timescaledb_internal._hyper_1_2_chunk, _timescaledb_internal._hyper_1_3_chunk}');
  

• Merge two s concurrently, allowing reads:

  CALL merge_chunks('_timescaledb_internal._hyper_1_1_chunk', '_timescaledb_internal._hyper_1_2_chunk', concurrently => true);
  

Arguments

The syntax is:

-- Merge two chunks:
CALL merge_chunks(
    chunk1 = '<chunk_name_1>',
    chunk2 = '<chunk_name_2>',
    concurrently = true | false
);

-- Merge multiple chunks:
CALL merge_chunks(
    chunks = ARRAY['<chunk_name_1>', '<chunk_name_2>', ...]::REGCLASS[]
);

-- Merge multiple chunks concurrently:
CALL merge_chunks_concurrently(
    chunks = ARRAY['<chunk_name_1>', '<chunk_name_2>', ...]::REGCLASS[]
);

You can merge either two s, or an arbitrary number of s specified as an array of identifiers. When you call merge_chunks, you must specify either chunk1 and chunk2, or chunks. You cannot use both arguments.

The concurrently option is only available for the two-chunk overload. To merge an array of s concurrently, use merge_chunks_concurrently() instead.

Name	Type	Default	Required	Description
chunk1, chunk2	REGCLASS	-	✖	The two s to merge in partition order
chunks	REGCLASS[]	-	✖	The array of s to merge in partition order
concurrently	BOOLEAN	false	✖	Set to true to allow reads on the merging s. Only available for the two-chunk overload.

Returns

This procedure does not return a value. Upon successful completion, the specified s are merged into a single .