It seems that for some reason the mechanism that should avoid saving a combination to the same
reporter/partner/year/item/flow/data_type combination is not working properly. Indeed on 2018-01-30 I found that the corrections file had 28 combinations (of which 26 by the same user) that were corrected twice out of a total of 2810 corrections. Probably, the most important information is that of these 28 combinations, 25 were due to the "Multiple corrections" feature.
It seems that for some reason the mechanism that should avoid saving a combination to the same
reporter/partner/year/item/flow/data_type
combination is not working properly. Indeed on 2018-01-30 I found that the corrections file had 28 combinations (of which 26 by the same user) that were corrected twice out of a total of 2810 corrections. Probably, the most important information is that of these 28 combinations, 25 were due to the "Multiple corrections" feature.