Issues with DCI's,transforms, and SQL errors.

Started by cbwecomm, July 29, 2025, 08:35:34 AM

Previous topic - Next topic

cbwecomm

I have a new 5.2.4 install, and I was just trying to get some of the initial stuff set up.  I was building some dashboards, and one of the DCI's polling from SNMP was giving me a value that needed to be /10  (It's a voltage, and the data polled from SNMP is like 578, when the voltage should be 57.8).    So, I go to the template to add the transform, and I cant get it to work (even though the test works fine).  I mess with it for a while, and then by chance I notice in the logs that there is an error every time I "start data recalculation" on the node's DCI.    I'm not 100% sure what I'm reading, but it seems that there is a column missing. 

Database query failed (Query: SELECT date_part(\u0027epoch\u0027,idata_timestamp)::int,raw_value FROM idata_sc_30 WHERE node_id\u003d245 AND item_id\u003d1094 ORDER BY idata_timestamp; Error: 42703 ERROR: column \"node_id\" does not exist\nLINE 1: ..._timestamp)::int,raw_value FROM idata_sc_30 WHERE node_id\u003d24...\n ^)"

I figured it was something that I was doing, but then I also noticed that 5.2.3 listed "#140 (SQL errors after converting database to TimescaleDB)" ...and this was installed as a TimescaleDB...  So, I'm thinking that maybe this is jut another, or unfixed bug in 5.2.4.   It was a new install of 5.2.4 though, so no conversions, or upgrades. 

Not sure if anyone can confirm, or point me in the right direction?   

EDIT:  I forgot to mention, that nxdbmgr check passes with no errors. 

Thanks!

Filipp Sudanov

#140 is relate to new nxdbmgr functionality - it now can do in-place conversion of Postgres DB to Timescale. 

But this here is a different bug - I've created a ticket for it: https://track.radensolutions.com/issue/NX-2819
This issue affects only recalculation of historic values, but new values should be transformed correctly. 

cbwecomm

Thank you Filipp!   

I believe your assessment is correct; eventually, I noticed (later that night) that the transform was working correctly.  Since I was forcing the recalculation, getting the error, and noting that the value wasn't changing, I thought it was broken entirely.  

I've cleared the historic data, and now the data/graph is displaying correctly.  

Appreciate the help and fast response as always!