Open rorads opened 8 years ago
The problem is for the new activities these problems shouldn't occur. These errors are from the previous IATI version data for e.g. format was not mandatory in v1.03 but is now required, so the legacy dpcument-link data (without format attribute) gets into the xml. It would definitely be good to catch them at AidStream end itself, validate it against the schema before publishing the file. Lets discuss.
4 different issues found with 4 different organisations:
Emmanuel: missing category code in document-link element of Organisation data file (fixed) Feed the Minds: AidStream had generated an empty xml file (fixed by republishing) International Medical Corps: missing name element in Organisation data file (fixed) Africa Educational Trust: missing dates in transactions (ticket sent to organisation)
Hello Everyone,
Having just completed an audit of publishers who've recently upgraded to 2.01, and publishers that have files failing schema validation, I've noticed a correlation.
Each of the following publishers are currently failing validation on one or more of their files, and each have very recently upgraded to 2.01 via Aidstream (within the last 4 days, according to the Aidstream Twitter feed).
There are others that upgraded earlier, but I'm yet to collate their data quality figures. Regardless, there was a similar increase in both 2.01 numbers and schema validation errors in the first week of the 2.01 upgrade.
The issue with almost all of these is that there is an element that is unexpected or missing, though they aren't systematic enough to be able to pin one down. Document links seem to be a fairly large proportion, but only form a slight majority.
Am I correct in thinking that Aidstream should either catch these validation errors? Do you think someone could look into this?
Many thanks, and kind regards, Rory