GN 4 unable to harvest Isogeo catalogue (CSW endpoint)

GeoNetwork GeoNetwork User
1

Hi,

I try to harvest an Isogeo catalogue from GeoNetwork 4.4.9 using CSW.

The records of this catalogue is accessible there:

https://open.isogeo.com/s/715fcd2e79f24bf4b72af6fdd4cb7acc/XqIpKIODFa_pc-eq5J9oWOWioyxu0

The GetCapabilities URL of the CSW endpoint of the Isogeo instance:

https://services.api.isogeo.com/ows/s/715fcd2e79f24bf4b72af6fdd4cb7acc/c/e7cae702d088465cb72bd71556450e7e/XqIpKIODFa_pc-eq5J9oWOWioyxu0?service=CSW&request=GetCapabilities

The GetRecords POST endpoint is:

https://services.api.isogeo.com:443/csw/s/715fcd2e79f24bf4b72af6fdd4cb7acc/c/e7cae702d088465cb72bd71556450e7e/XqIpKIODFa_pc-eq5J9oWOWioyxu0/GetRecords

I have configured the harvester with the outputSchema “http://www.isotc211.org/2005/gmd”.

The harvesting result seems good: “19 record(s) harvested in 26 seconds”

But only one record is shown in the list of havested records.

I get the same strange results with a production GeoNetwork (4.4.9) instance and a dockerised instance on my laptop (4.4.6). I get the same kind of result with another Isogeo catalogue.

The GeoNetwork logs do not mention any error/warning at all:

geonetwork-1               | 2026-06-23T09:20:31,441 INFO  [geonetwork.harvester] - Starting harvesting of test isogeo eptb sgl
geonetwork-1               | 2026-06-23T09:20:31,477 INFO  [geonetwork.harvester] - Started harvesting from node : test isogeo eptb sgl (CswHarvester)
geonetwork-1               | 2026-06-23T09:20:35,747 WARN  [geonetwork.harvester] - Declared number of returned records (19) does not match requested record count (20)
geonetwork-1               | 2026-06-23T09:20:35,747 INFO  [geonetwork.harvester] - Total records processed in all searches :19
geonetwork-1               | 2026-06-23T09:20:35,747 INFO  [geonetwork.harvester] - Ended harvesting from node : test isogeo eptb sgl (CswHarvester)

[owslib retrieves the metadata records from this catalog with ease.]

Do you have any idea of what is going wrong during the harvesting process of this catalogue?

Thank you for your help?

2

Hi,

the first thing to do is probably to check in the admin dashboard how many records are indexed vs. how many are in the database. It sounds like most records might fail indexing for some reason, which would explain their absence in the search results.

Cheers,

Olivia

3

Thank you for this tip.

The total number of records and number of indexed records are equal on my local dockerised GeoNetwork. But the admin dashboard indicates that all the metadata records harvested from these Isogeo catalogs encountered an indexing error:

[1:2184] failed to parse field [revisionDateForResource] of type [date] in document with id 'urn:isogeo:metadata[…]’

I have to look deeper in the input metadata records…

Thank you a lot.