Timescaledb: `TRUNCATE TABLE` deletes data but does not remove underlying chunks

[abrkn]

I experienced some odd behavior after running TRUNCATE TABLE on a table that has a hyper table. I didn't have time to investigate at the time and re-created the table.

[olofr]

Hi, thanks for reporting this. When I truncate an hypertable, the existing chunks are not removed although they are empty. Is this the behavior that you saw and found odd ? I think removing the empty chunks on truncate would be a better behavior and will look into fixing it.

[mfreed]

Hi @abrkn thanks for this feedback.

We'll look into more expected behavior for TRUNCATE in the future (i.e., delete the existing chunks), but for now, you can turn to our drop_chunks interface for the expected behavior:

 SELECT drop_chunks(NOW() + interval ‘1 year’, tablename)

drop_chunks deletes all chunks if all of their data are beyond the cut-off point (based on chunk constraints). Here I'm just pushing out the leading edge by 1 year to just mean something "way in the future". Not the prettiest interface, but can give you the functionality you want for now.

[mfreed]

Also, my testing confirms the same as @olofr : after TRUNCATE, all the data is deleted, just that the existing chunks remain (but now are empty).

[mfreed]

Just a head's up: PR currently being reviewed: https://github.com/timescale/timescaledb/pull/131

[mfreed]

Fixed as part of 0.3.0 release (by @olofr).

https://github.com/timescale/timescaledb/releases/tag/0.3.0

[cercata]

Hi @abrkn thanks for this feedback.

We'll look into more expected behavior for TRUNCATE in the future (i.e., delete the existing chunks), but for now, you can turn to our drop_chunks interface for the expected behavior:

 SELECT drop_chunks(NOW() + interval ‘1 year’, tablename)

drop_chunks deletes all chunks if all of their data are beyond the cut-off point (based on chunk constraints). Here I'm just pushing out the leading edge by 1 year to just mean something "way in the future". Not the prettiest interface, but can give you the functionality you want for now.

This deletes all chunks ... I guess the drop_chunks() function has evolved since you wrote this, and the behaviour has changed

[LRagji]

Just an update truncate table cascade fails with default config it is asking to increase the shared memory limit..