You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
There's already a PR made by Vincent that addresses this at Schema level. Worth verifying the pull request, removing the rule/test and keep it at schema level.
To confirm with @IATI/bas that this is not a problem to enforce at schema level as well (adding a hard 0-100 schema check would not help values "close enough to be rounded" such as 99.8 so might put a lot of publishers off)
The text was updated successfully, but these errors were encountered:
EDIT: this is labeled as 2.03 only because in 2.01 and 2.02 we only check the percentage value to be positive (>= 0.0). The assumptions below are for 2.03 only
This is currently undergoing internal discussion with @IATI/bas . Main takeaway points:
if we enforce this at Schema level (where, correctly assumed already, makes sense to live), it'll enforce a hard check on the ranges potentially risking data being excluded from the datastore; this is the option I'm more keen on, but I also understand this needs to go through a bug fix proposal
if we keep it at Ruleset level, then the rule is correct and won't prevent data from being imported; validation scores will be affected obviously but gives publishers space to fix it; my least preferred option, but time/effort wise would not require any further action from us
I guess this is a standard management question, whether percentage bounds should be schema-enforced. But doing so would probably require an integer upgrade as it's not been a part of the standard for so long.
IATI-Rulesets/rulesets/standard.json
Lines 109 to 132 in b88c627
There's already a PR made by Vincent that addresses this at Schema level. Worth verifying the pull request, removing the rule/test and keep it at schema level.
To confirm with @IATI/bas that this is not a problem to enforce at schema level as well (adding a hard 0-100 schema check would not help values "close enough to be rounded" such as 99.8 so might put a lot of publishers off)
The text was updated successfully, but these errors were encountered: