Extract values from a column in a Dataframe based in starting letters

Question

Hope everyone is doing well. I am using pandas and numpy and I would like to extract column values based on the first 3 letters (ap.) from a Dataframe. Below is an example of my Dataframe.

Name	Number
Orange	2
APple	6
Ap.ricot	1
AP.19	1
Juap.rte	3

I've tried df[df['Name'].str.lower().str.contains('ap.', na=False)].Name.unique() but it does not fully do the trick.

Output:

['AP.19','Ap.ricot']

The output should ideally be a list that I can then save onto a variable. Additionally, the 3 letters need to be at the start and in this order.

I am very new to Python so please explain as clearly as possible. Thank you.

Do you want to take out values where your code finds 'Ap' only? — Prateek Jain, Dec 22 '20 at 17:52
No, I would like to extract the values in a list form, which start with 'ap.' Of course, would also need to convert everything to lower case, but the output so be in its original form. So exactly the output I have, otherwise have 'Ap' would also include 'APple'. — Ammar Kamran, Dec 22 '20 at 17:54

Celius Stingher · Answer 1 · 2020-12-22T18:02:07.303

2

Given the comments in the post, I believe you can get it done with:

ap = [x for x in df['Name'] if x.lower().startswith('ap.')]

And if you wish to not have duplicates, then you can use:

ap = [x for x in df['Name'].unique() if x.lower().startswith('ap.')]

edited Dec 22 '20 at 18:02

answered Dec 22 '20 at 17:56

Celius Stingher

17,835
6
23
53

1

Thank you @Celius Stingher That is exactly what I was looking for. Wish you a nice day. – Ammar Kamran Dec 22 '20 at 18:01
1

Have a great day and Merry Christmas! Feel free to mark my answer as accepted so we know the problem has been solved! – Celius Stingher Dec 22 '20 at 18:01

score 1 · Answer 2 · answered Dec 22 '20 at 17:59

1

This may help you:

final = []


df['NameCopy'] = df['Name'].str.lower()


for index,row in df.iterrows():
   if row['NameCopy'].find('ap.') != -1:
      final += [row['Name']]
   else:
      pass

print(final)

answered Dec 22 '20 at 17:59

Prateek Jain

231
1
11

Thank you! This also works. I will accept and approve the answer as soon as StackOverflow allows me. – Ammar Kamran Dec 22 '20 at 18:04
2

This is very inefficient as it is looping over a dataframe and should be discouraged. – Celius Stingher Dec 22 '20 at 18:06
1

Yes agree to @CeliusStingher: One should avoid iterating. https://stackoverflow.com/a/55557758/6660373 btw merry Christmas to all of you in advance :) – Pygirl Dec 22 '20 at 18:07
@CeliusStingher: Good point, but such loops are better where there are more conditions you need to loop through within the same iteration. Which is why I prefer using these in my datasets. – Prateek Jain Dec 22 '20 at 18:08
You should avoid that. Visit the link that I have mentioned. It will help you in future :) – Pygirl Dec 22 '20 at 18:09
2

There is no conditions which you need to meet, and it has been proven many times there are vectorized ways of dealing with conditions such as `np.where()`. Using for loops should be avoided and discouraged when dealing with pandas dfs – Celius Stingher Dec 22 '20 at 18:09

score 1 · Accepted Answer · answered Dec 22 '20 at 18:03

1

try:

df[df['Name'].str.match('^(ap[.])', case=False)].Name.unique()

array(['Ap.ricot', 'AP.19'], dtype=object)

answered Dec 22 '20 at 18:03

Pygirl

12,969
5
30
43

1

I like the use of regex here :) Merry Xmas :D +5 – Celius Stingher Dec 22 '20 at 18:09

Extract values from a column in a Dataframe based in starting letters

3 Answers3