CODEWITHSUNDEEP
python-3.xpandas
Agniv
Agniv

Reputation: 137

Pandas Df logical expression for Not In

I am converting java code into a python for re-writing an legacy etl application. In my legacy application the logical expression for data filtration is like below:

// code sample : use input_row to define the condition. // input_row.columnName1.equals("foo") ||!(input_row.columnName2.equals("bar"))

(!(row3.one_GRP_NM == null && row3.two_GRP_NM == null) && 
 !(row3.one_GRP_NM != null && row3.one_GRP_NM.equals(row3.two_GRP_NM)))

Now I converted into below -

df[((((df['one_grp_nm']=='null') & (df['two_grp_nm']=='null')) &
     ((df['one_grp_nm']!='null') & (df['one_grp_nm']==df['two_grp_nm'])))]

But I am not able to figure it out how to convert the ! expression at the beginning of the logic. Could you please guide how to implement not matching/not in in pandas df.

Upvotes: 0

Views: 80

Answers (2)

Agniv
Agniv

Reputation: 137

thanks lot for your help. Looks like I need to modify little bit for the null column as well -> 

df[(~((df['one_grp_nm'].isnull()) & (df['two_grp_nm'].isnull())) &\
    ~(((df['one_grp_nm'].notna()) & (df['one_grp_nm']==df['two_grp_nm'])))
  )]

Upvotes: 1

Toukenize
Toukenize

Reputation: 1420

In Pandas, negation (i.e. not) is denoted using a ~

For your application, it should be :

df[((~((df['one_grp_nm']=='null') & (df['two_grp_nm']=='null')) &\
     ~((df['one_grp_nm']!='null') & (df['one_grp_nm']==df['two_grp_nm']))))]

Note that if you are breaking across line after &, you should add a backslash \ otherwise it will be a syntax error.

Another example on not in in Pandas:

match = [1,2,3]
df = pd.DataFrame({'col_1':[1,2,3,4,5,6]})

# Note the ~
# You can read it as `not (isin)`
df[~df['col_1'].isin(match)]

will give you:

    col_1
3   4
4   5
5   6

Upvotes: 1

Related Questions