Apply lambda on a column in pandas

Question

I have the below data frame

ipdb> csv_data
  country_edited  sale_edited  date_edited  transformation_edited
0          India       403171     21091956                      1
1         Bhutan       394096     21091956                      2
2          Nepal       361372     21091956                      3
3         madhya       355883     21091956                      4
4          sudan       262949     21091956                      5

and below is my code

transfactor_count = 5.6
csv_data["transformation_edited"] = csv_data["transformation_edited"].apply(lambda x: x * transfactor_count)

But the above code was giving me an error

*** NameError: global name 'transfactor_count' is not defined

How to solve this ?

Actual code

for foreign_key in data_mapping_record.csvdatabasecolumnmapping_set.all():
    data_type = foreign_key.data_type
    trans_factor = foreign_key.tranformation_factor
    if data_type == "Decimal":
        import ipdb; ipdb.set_trace()
        csv_data[foreign_key.table_column_name] = csv_data[foreign_key.table_column_name].apply(lambda x: x * trans_factor )
    elif data_type in ["Date", "Datetime"]:
        csv_data[foreign_key.table_column_name] = csv_data[foreign_key.table_column_name].apply( lambda d: datetime.strptime(d, dates[date]) )

For me it works, but why dont use `csv_data["transformation_edited"] = csv_data["transformation_edited"] * transfactor_count` or `csv_data["transformation_edited"] = csv_data["transformation_edited"].mul(transfactor_count)` ? — jezrael, Jan 25 '17 at 10:00
basically i was using django and i am gonna get transfactor_count from database — Shiva Krishna Bavandla, Jan 25 '17 at 10:03
Is possible use `csv_data["transformation_edited"].mul(foreign_key.tranformation_factor)` ? — jezrael, Jan 25 '17 at 10:06
like `csv_data[foreign_key.table_column_name] = csv_data[foreign_key.table_column_name].mul(foreign_key.tranformat‌ion_factor)` — jezrael, Jan 25 '17 at 10:08
*** TypeError: ufunc 'multiply' did not contain a loop with signature matching types dtype(' — Shiva Krishna Bavandla, Jan 25 '17 at 10:10
It seems more complicated, how works csv_data[foreign_key.table_column_name] = csv_data[foreign_key.table_column_name]* foreign_key.tranformation_factor) ? — jezrael, Jan 25 '17 at 10:19
yeah it worked i need to convert `foreign_key.tranformation_factor` to `int(foreign_key.tranformation_factor)`, but the problem is why can't i access the variables just before i declared ? like for example i have declared `dates = { 'dd/mm/yy': "%d/%m/%Y", 'mm/dd/yy': "%m/%d/%Y", }` as a dictionary, but when i tried to access like `csv_data[foreign_key.table_column_name].apply( lambda d: datetime.datetime.strptime(d, dates[foreign_key.date_format]) )` it says `*** NameError: global name 'dates' is not defined` ? — Shiva Krishna Bavandla, Jan 25 '17 at 10:21
Let us [continue this discussion in chat](http://chat.stackoverflow.com/rooms/133966/discussion-between-shiva-krishna-and-jezrael). — Shiva Krishna Bavandla, Jan 25 '17 at 10:28

score 4 · Accepted Answer · answered Jan 25 '17 at 12:09

As the error already says, the lambda function can't find the global variable. You can try make it global by:

global transfactor_count
transfactor_count = 5.6
csv_data["transformation_edited"] = csv_data["transformation_edited"].apply(lambda x: x * transfactor_count)

But as jezrael pointed out:

csv_data["transformation_edited"] = csv_data["transformation_edited"] * transfactor_count

is much more elegant.

you might have to change the datatype to float before:

csv_data["transformation_edited"] = csv_data["transformation_edited"].astype(float) * transfactor_count

score -1 · Answer 2 · edited Dec 03 '18 at 11:48

-1

You can use lambda on dataframe as below:

transfactor_count = 5.6;
csv_data['transformation_edited']=map(lambda x: x * transfactor_count, csv_data['transformation_edited'])

edited Dec 03 '18 at 11:48

Rahul Agarwal

4,034
7
27
51

answered Jan 25 '17 at 10:13

sandeep

26
1

Apply lambda on a column in pandas

Actual code

2 Answers2