Manufacturing Batch Data

[a-dub]

Philosophical question for folks. 

When recording historical data for a manufacturing batch (for retainment in a "batch report" (uptime, average temp, etc)) what thingworx data storage method would be best?

After looking into everything I would think that--for example--temperature data could be recorded in a stream (as it is time series) and tagged with the batch unique identifying number.

however, where would other data be captured which isn't time series. For example--Uptime, or units produced, or yield.

Would it be best practice for each manufacturing batch to be created as it's own "thing" with specific persistant properties holding that data.

Or would it be more logical to have a "batches" table which has an uptime/units produced/yield column?

Would it be better to have each batch identification number a "tag" and then have an uptime data table where each row is tagged with that batch identification number? 

I guess what I'm getting at is: should each batch be its own: (1) thing, (2) row in data table, or (3) tag in many data tables? 

Is there some other better method that I'm missing?

[PaiChung · Accepted solution]

Some of this depends on how you will use the data going forward.

If this is going to be a lot of data mostly for historical analysis and insight I would recommend a properly indexed separate database

If this is short term heavily queried (lot of filtering), probably a DataTable

if this is short term, medium term, with mainly timestamp querying a Stream

You can also go with a hybrid approach, short / medium term (under 500K records) with reasonable usage in a Stream and then offload to a separate database for historical analysis of the long term data.

There are some articles on  Scaling and DataTables vs. Streams that might be helpful to further your understanding.

[slangley]

Hi @a-dub.

If the previous response answered your question, please mark it as the Accepted Solution for the benefit of others with the same question.

Regards.

--Sharon