Timescaledb: Is it possible to create_hypertable() without table having column of type timestamp, Integer , DateTime? What is solution for this.

[vivekanandsh]

Relevant system information:

OS: Windows 10 x64, etc]
PostgreSQL version : 10.x
TimescaleDB version 1.1.4
Installation method: Installer

Describe the bug

Hello all,
How to partition table (in postgresqlDB) with table don't have column of type timestamp, Integer , DateTime integer ?

CREATE TABLE rep_backup
(
type character varying(100) COLLATE pg_catalog."default" NOT NULL,
id character varying(120) COLLATE pg_catalog."default" NOT NULL,
data bytea NOT NULL,
CONSTRAINT pk_replication_backups PRIMARY KEY (type, id)
)
above table don't have any column of type timestamp, Integer , DateTime.
Want to know for above table hash partition is possible , if yes please guide how to do?
Please provide sample example if possible.
Urgent one help to clarify/reply asap.

[mfreed]

I'm a little uncertain ... for what are you trying to use TimescaleDB? Is this somehow time-series data?

[vivekanandsh]

Mike, 
Thanks for quick reply. TimescaleDB want to use for PostgreSQLDB extension as we have more time related data and wan to partition. 
I have done for  partition for some tables already, only question is below table don't have column of type  timestamp, Integer , DateTime.
CREATE TABLE rep_backup
(
type character varying(100) COLLATE pg_catalog."default" NOT NULL,
id character varying(120) COLLATE pg_catalog."default" NOT NULL,
data bytea NOT NULL,
CONSTRAINT pk_replication_backups PRIMARY KEY (type, id)
)
So, want to know can i do HASH partition ? As per documentation hash partition used as multi dimension  partition _https://docs.timescale.com/latest/api#add_dimension_ .
Just help me clarify, MUST need column of type  timestamp, Integer , DateTime  in table to create hypertable or without those those types using hash partition possible.  

Regards, Vivek

[mfreed]

Hypertables currently require some type of timestamp/integer/date.

You might consider with the above is you need to partition at all (what benefit do you get from it?) or use native PG partitioning.

[vivekanandsh]

May i know the timeline for the timescaledb version compatible for PostgreSQL 12.x.
Hope many looking for this version. :)

[loringao]

We have similar user scenarios, but there are id(bigint) and create_time(timestamp), it's business transcation data(not a IoT data) , we expect get the stable &high throughput from timescale，does it reasonable?

[k-rus]

May i know the timeline for the timescaledb version compatible for PostgreSQL 12.x.
Hope many looking for this version. :)

See discussion #1456