Time Series query error with Redshift 'REAL' data type

[doanbutar]

What happened?
When trying to create a Time Series visualization with a Redshift data source that has a column with data type REAL, the query fails with an HTTP 500 - Internal Server Error.

The same query is running on the same column with int data type works.

What was expected to happen?
Queries on REAL data type columns should also complete successfully.

Steps to reproduce the problem:

1. Create a redshift cluster with a column with int data type
2. Create a Time Series dashboard in Grafana
3. Run a time series query that selects this int data type column
4. Create additional column in redshift with real data type, and copy over int data to the real data type column
5. Modify time series query to use real datatype column instead
6. Query fails with "Internal Server Error" (see attached screenshot)

Version numbers (grafana, prometheus, graphite, plugins, operating system, etc.):
Grafana version: v8.4.7
Redshift version: v1.3.0
OS: Amazon Linux 2

Configuration information:
Redshift table schema is also attached as screenshot.

[doanbutar]

//Redshift Int type

//Redshift listing table schema

//Redshift real type

[doanbutar]

Looking at this: #33
Inserting real datatype and also query seems to work fine.

[fridgepoet]

Thanks @doanbutar for the detailed bug report.

I was able to reproduce the issue by:

1. In AWS, create a redshift cluster
2. Create a table with a column type real: create table foo(test real);
3. Insert 1 value into the real column: insert into foo(test) values (1.111111);
4. In Grafana, create a time series panel and execute the query select * from foo

Result: error message Query data error

(confirmed on same version as you and on latest Grafana 9.4.0-pre with Redshift 1.5.0)

[grafpat]

Hi @fridgepoet , any update here? I see that this has been moved to Backlog, is there any update on rough timeline that this will be worked on?

[fridgepoet]

Hi @grafpat we don't have a firm timeline yet, will try to keep you informed when we pick this up.

[fridgepoet]

For future readers: For testing this out, it's much easier to write a query like SELECT CAST('123456' AS REAL); than to create a cluster/table/column of a certain type.