Need to search a string column values from a list of strings. The strings in the search list are only a substring of the values in the column
df = pd.DataFrame(data={'text':['abc def', 'def ghi', 'poi opo', 'aswwf', 'abcs sd'], 'id':[1, 2, 3, 4, 5]})
Out [1]:
text id
0 abc def 1
1 def ghi 2
2 poi opo 3
3 aswwf 4
4 abcs sd 5
search = ['abc', 'poi']
Required:
Out [2]:
text id
0 abc def 1
1 poi opo 3
2 abcs sd 5
Use Series.str.contains with boolean indexing - all values of list are joined by | for regex OR:
pat = '|'.join(search)
df1 = df[df['text'].str.contains(pat)]
print (df1)
text id
0 abc def 1
2 poi opo 3
4 abcs sd 5
@jezrael'answer is great, provided the patterns to search contain no special characters like |. But you can process every element at a time and do a global or at the end. If you want to search strings containing special characters, you can use:
df[pd.concat([df.text.str.contains(i, regex=False) for i in search], axis=1).any(axis=1)]
it gives as expected:
text id
0 abc def 1
2 poi opo 3
4 abcs sd 5
