The count for the null entry matches the total count, even though there are 8,064 non-null entries. We don't expect to observe null and non-null values in the same project, so this looks like a problem with how the null count is calculated.
$ curl 'https://service.azul.data.humancellatlas.org/index/files?catalog=dcp39' | jq '.termFacets.tissueAtlas'
% Total % Received % Xferd Average Speed Time Time Time Current
Dload Upload Total Spent Left Speed
100 553k 100 553k 0 0 63921 0 0:00:08 0:00:08 --:--:-- 127k
{
"terms": [
{
"term": "Lung",
"count": 6264
},
{
"term": "Retina",
"count": 1050
},
{
"term": "Blood",
"count": 323
},
{
"term": "Kidney",
"count": 237
},
{
"term": "Gut",
"count": 190
},
{
"term": null,
"count": 520065
}
],
"total": 520065,
"type": "terms"
}
The count for the
null
entry matches the total count, even though there are 8,064 non-null entries. We don't expect to observenull
and non-null
values in the same project, so this looks like a problem with how thenull
count is calculated.