I have a df called df like so. The tag_position is either a string or list. but I want them to be all strings. How can i do this? I also want to remove the white space at the end.
input
id tag_positions
1 center
2 right
3 ['left']
4 ['center ']
5 [' left']
6 ['right']
7 left
expected output
id tag_positions
1 center
2 right
3 left
4 center
5 left
6 right
7 left
You can explode and then strip:
df.tag_positions = df.tag_positions.explode().str.strip()
to get
id tag_positions
0 1 center
1 2 right
2 3 left
3 4 center
4 5 left
5 6 right
6 7 left
You can join:
df['tag_positions'].map(''.join)
Or:
df['tag_positions'].str.join('')
Try with str chain with np.where
df['tag_positions'] = np.where(df['tag_positions'].map(lambda x : type(x).__name__)=='list',df['tag_positions'].str[0],df['tag_positions'])
Also my favorite explode
df = df.explode('tag_positions')
you can convert with apply method like this
df.tag_positions = df.tag_positions.apply(lambda x : ''.join(x) if type(x) == list else x)
if all the lists have a length of 1 you can do this also:
df.tag_positions = df.tag_positions.apply(lambda x : x[0] if type(x) == list else x)
You can use apply and check if an item is an instance of a list, if yes, take the first element. and then you can just use str.strip to strip off the unwanted spaces.
df['tag_positions'].apply(lambda x: x[0] if isinstance(x, list) else x).str.strip()
OUTPUT
Out[42]:
0 center
1 right
2 left
3 center
4 left
5 right
6 left
Name: 0, dtype: object
