Authorities name analysis: Error in recordcount?

[jaFHP]

While performing the authorities name analysis I noticed that in the output file authorities-by-schema.csv

recordscount = instancecount,

even though other results (e.g. in output file authorities-by-categories.csv) suggest, that there should be more instances than records.

I analysed 2 different catalogues with QA-Catalogue Version 0.7.0, both in marcxml, one with a local version of marc, one without, both showed this issue. I have not yet been able to find the possible error in the code, but it looks to me as if the instancecountis written into authorities-by-schema.csv as both instancecount and recordcount.

Example

from output file authorities-by-categories.csv:

category	recordcount	instancecount
Personal names	1118033	1844024
Corporate names	49418	54339
Meeting names	10	11
Titles	61445	62714
SUM	1228906	1961088

from output file authorities-by-scheme.csv:

id	field	location	recordcount	instancecount
1	100	$2	939644	939644
4	110	$2	23695	23695
10	111	$2	3	3
3	130	$2	21404	21404
2	700	$2	870063	870063
5	710	$2	30094	30094
9	711	$2	8	8
6	730	$2	855	855
8	800	$2	34317	34317
11	810	$2	550	550
7	830	$2	40455	40455
		SUM	1961088	1961088

[pkiraly]

@jaFHP However 0.7.0 is the latest published version, it is rather old, and it might contain bugs that is already fixed. To reveal the problem I see two options:

1. you can try to use the latest version, which is available in a dockerized version. Here you can find the instructions how to install it: https://github.com/pkiraly/qa-catalogue/wiki/D%E2%80%90A%E2%80%90CH-Koha-Hackfest-Workshop-(G%C3%B6ttingen,-2025%E2%80%9010%E2%80%9009)
2. you enable me to download your dataset, and I can run the analysis (if you do not share a URL, please write me an email to pkiraly at gwdg punkt de).

[jaFHP]

Thank you @pkiraly !
I tried the dockerized version with the instructions you linked, because i cannot share the data i used unfortunately.. The output for the authorities name analysis shows the exact same problem as described above.

I also tried to look at other interactive reports online. KBR seems to show the same problem (probably also not using the newest version of qa-catalogue, but to illustrate:

)

Am I missing something? Can instances==records in every case, even when sums above suggests otherwise?