CODEWITHSUNDEEP
pythonpandas
richie
richie

Reputation: 18638

TypeError: Could not compare <type 'str'> type with Series

I get the following error when I try to filter a Dataframe by using a string

TypeError: Could not compare <type 'str'> type with Series

This is my code;

data = pd.read_csv('data.csv')
fildata = data[(data['cat1'] == 'FALSE') & (data['cat2'] != '') & (data['cat3'] == 'FALSE')]

EDIT 1:

Here's how the data looks like;

count,word,cat1,cat2,cat3
1021,.,FALSE,,FALSE
825,the,TRUE,the,FALSE
693,and,TRUE,and,FALSE
647,of,TRUE,of,FALSE
646,",",FALSE,,FALSE
435,to,TRUE,to,FALSE
353,will,TRUE,will,FALSE
297,in,TRUE,in,FALSE
274,be,TRUE,be,FALSE

EDIT 2:

But why does this work?

data1 = pd.DataFrame({'cat1':[1,2,3,4],'cat2':[2,3,1,4],'cat3':[3,1,2,4]})

fildata = data1[(data1['cat1'] == 1) & (data1['cat2'] != 0) & (data1['cat3']== 3)]

This results in;

cat1 cat2 cat3 0 1 2 3

EDIT 3:

I guess the problem is with the type. 'cat1' & 'cat2' are of the type 'bool'

Upvotes: 2

Views: 3178

Answers (1)

EdChum
EdChum

Reputation: 394021

The following worked for me:

In [114]:

temp = """count,word,cat1,cat2,cat3
1021,.,FALSE,,FALSE
825,the,TRUE,the,FALSE
693,and,TRUE,and,FALSE
647,of,TRUE,of,FALSE
646,",",FALSE,,FALSE
435,to,TRUE,to,FALSE
353,will,TRUE,will,FALSE
297,in,TRUE,in,FALSE
274,be,TRUE,be,FALSE"""
data = pd.read_csv(io.StringIO(temp))
fildata = data[(data['cat1'] == False) & (data['cat2'].isnull() ) & (data['cat3'] == False)]
In [115]:

fildata
Out[115]:
   count word   cat1 cat2   cat3
0   1021    .  False  NaN  False
4    646    ,  False  NaN  False

[2 rows x 5 columns]

The problem you have is that the string FALSE/TRUE are boolean dtypes as interpreted by read_csv:

In [112]:

data.dtypes
Out[112]:
count     int64
word     object
cat1       bool
cat2     object
cat3       bool
dtype: object

so your comparison should be against this type and not the string 'FALSE'

Upvotes: 3

Related Questions