group date value
a 2020-06-01 1
a 2020-06-02 1
a 2020-06-03 -1
b 2020-06-01 2
b 2020-06-02 -1
b 2020-06-03 -1
c 2020-06-02 -3
c 2020-06-03 3
c 2020-06-04 2

For each date combination I want to find the groups that the value's sum is equal to zero.
Note that I don't want to sum values across different groups.
For exemple, using this dataframe, I would like an output that returns me this information:

2020-06-01 and 2020-06-03: a
2020-06-02 and 2020-06-03: a, c
2020-06-01 and 2020-06-02 and 2020-06-03: b

hmmm possibly needs more cleaning up as C is satisfied for 2020-06-02 & 2020-06-03 and
there is no value for C for 2020-06-01 which falsely allows 2020-06-01 & 2020-06-02 & 2020-06-03 to be rated for c