删除重复但优先保留在熊猫中

Question

Here is a df:

COL1 COL2 COL3 
seqA NA 10
seqA Unknown 5
seqA Cow 50
seqB NA 2
seqC NA 2
seqC Unknown 2
seqC Bird 6
seqC Cow 1
seqD Unknown 30
seqD Shark 2

so the idea would bee to remove duplicated COL1 value and keep only one with the lowest COL3 BUT only take ones with NA or Unknown containt if there is no other COL3 value < 10

for instance for SeqA

I keep

seqA Unknown 5

because thise one is > 10 :

seqA Cow 50

but in seqC I keep :

seqC Cow 1

because it is <10

In the exemple the expected output would be :

COL1 COL2 COL3 
seqA Unknown 5
seqB NA 2
seqC Cow 1
seqD Shark 2

So one idea would be to first do a

tab=df.sort_values(by=['COL3'], ascending = True)

But I do not know how to integrate the priority by the fact that everything different from Unknwown or NA is a priority except it its COL3 > 10

Answer 1

Let us do filter then sort_values + drop_duplicates

out = df[df.COL3.lt(10) | df.COL2.eq('Unknown')].sort_values('COL3').drop_duplicates('COL1').sort_index()
Out[47]: 
   COL1     COL2  COL3
1  seqA  Unknown     5
3  seqB      NaN     2
7  seqC      Cow     1
9  seqD    Shark     2