Skip to main content
Sets the chunk_time_interval on a . The new interval is used when new s are created, and time intervals on existing s are not changed.

Samples

For a TIMESTAMP column, set chunk_time_interval to 24 hours:
SELECT set_chunk_time_interval('conditions', INTERVAL '24 hours');
SELECT set_chunk_time_interval('conditions', 86400000000);
For a time column expressed as the number of milliseconds since the UNIX epoch, set chunk_time_interval to 24 hours:
SELECT set_chunk_time_interval('conditions', 86400000);

Arguments

NameTypeDefaultRequiredDescription
hypertableREGCLASS- or to update interval for.
chunk_time_intervalSee note-Event time that each new covers.
dimension_nameREGCLASS-The name of the time dimension to set the number of partitions for. Only use dimension_name when your has multiple time dimensions.
If you change time interval you may see a that is smaller than the new interval. For example, if you have two 7-day s that cover 14 days, then change chunk_time_interval to 3 days, you may end up with a transition covering one day. This happens because the start and end of the new is calculated based on dividing the timeline by the chunk_time_interval starting at epoch 0. This leads to the following s [0, 3), [3, 6), [6, 9), [9, 12), [12, 15), [15, 18) and so on. The two 7-day s covered data up to day 14: [0, 7), [8, 14), so the 3-day for [12, 15) is reduced to a one day . The following [15, 18) is created as a full 3 day . The valid types for the chunk_time_interval depend on the type used for the time column:
time column typechunk_time_interval typeTime unit
TIMESTAMPINTERVALdays, hours, minutes, etc
INTEGER or BIGINTmicroseconds
TIMESTAMPTZINTERVALdays, hours, minutes, etc
INTEGER or BIGINTmicroseconds
DATEINTERVALdays, hours, minutes, etc
INTEGER or BIGINTmicroseconds
SMALLINTSMALLINTThe same time unit as the time column
INTINTThe same time unit as the time column
BIGINTBIGINTThe same time unit as the time column
For more information, see partitioning.