Unable to sort by sum with limit

Welcome & Getting Started
1

I’m trying to query the top 5 project with highest usage.

My query is :

from (bucket: "test_repo")
|> range(start: -10d)
|> drop("columns": ["user", "version", "resource", "_stop", "_start"])
|> filter(fn: (r) =>
    r._measurement == "image_pull" and
    r._field == "cnt"
)
|> sum()
|> sort("columns" : ["_value"])
|> limit(n : 5)

Last 3 rows of the result returned by influx query is :

Table: keys: [_field, _measurement, project]
         _field:string     _measurement:string          project:string                  _value:int
----------------------  ----------------------  ----------------------  --------------------------
                   cnt              image_pull                zbentxoml                         49
Table: keys: [_field, _measurement, project]
         _field:string     _measurement:string          project:string                  _value:int
----------------------  ----------------------  ----------------------  --------------------------
                   cnt              image_pull             zedge-devid                        4283
Table: keys: [_field, _measurement, project]
         _field:string     _measurement:string          project:string                  _value:int
----------------------  ----------------------  ----------------------  --------------------------
                   cnt              image_pull                   zfaas                           4

It is still sorted by project name and more than 5 rows are returned.