Calculating pages classified

RillonDodgers

Hello!

I'm not positive this is a bug, so I wanted to post it here first just in case.

Here are two queries for calculating how many pages have been classified by the separate step

The first query returns 175 (I also did one where I joined the children nodes and the number was still 175)

The second query returns 134

It could be my misunderstanding of the "Pages Classified" custom stat, but I would assume it should be the same as calculating all of the pages in folders that have been assigned a content type.

This is the same for every batch I've tried it with and just to note these batches haven't had any human interaction. They're all fresh out of separate and ready for review

Does anyone have any insight into this?

Thanks!!

RillonDodgers

Here are the SQL queries for those interested. The ParentID of the first query is the batch root, or the parent folder of where all your documents get put under

-- Getting a total of how many pages have been classified by summing the
-- NumChildren column on folders that have been classified
select SUM(NumChildren)
from TreeNode
where ParentId = '72C4ED58-E3AF-4A5C-9888-8A35AECCF06C'
and json_value(Properties, '$.ContentTypeId') IS NOT NULL
and TypeName = 'Grooper.Core.BatchFolder'

-- Getting a total of how many pages have been classified by summing all the values
-- in CustomStats where the Name is Pages Classified
select SUM(cs.Value) from SessionStats ss
join CustomStats cs on cs.SessionId = ss.Id
where ss.BatchId = '0ee6691d-a729-45e9-a30c-df7b9bcd49d5'
and ss.StepName = 'Separate'
and cs.Name = 'Pages Classified'

jsprague

In Design Studio how many classified pages for that batch are there actually? Is it 175? And what node has that Id that is called in the first query? Is it the batch folder?

RillonDodgers

Yes, Design Studio shows 175, which matches the query.

I imagine in most people's use-case, it would be using the RootNodeId for the ParentId in the first query, but in our instance we put everything under a Level 1 folder. (we use it to store data for our ingress/egress operations)

Here is an example structure of our batches