使用 Python 中的循环根据字符串值修改新列中的行值

Question

I'd like to recode row values in a different column based on a string match in pandas using a loop. I found a way to do it by creating an entirely new column each time, but that doesn't work when I need to modify select rows from multiple columns at different points in the analysis.

Here was the solution I used with an example dataframe:

iris = sns.load_dataset('iris')
iris.head()
iris.species.value_counts()

pattern = ['setosa', 'virginica']
iris['new_column'] = 0
lis = []

for index, row in iris.iterrows():
  #print (row['species'])
  if any(ele in row.species for ele in pattern):
    lis.append('matched')
  else:
    lis.append("notmatched")

iris['new_column'] = lis

I know there may be other ways through list comprehensions in Pandas or using lambda/apply methods, but I'm requesting a solution using loops. (I don't have the full dataset here, but there's some complications with it and I believe a loop may be the most flexible).

Any suggestions on how to use a loop and string match to modify rows in a different column? Thank you and let me know if I can make this question better!

Answer 1

One of the simpliest loop solution is iterate by each value of column iris['species'] and append to list lis by condition with in :

pattern = ['setosa', 'virginica']
lis = []
for val in iris['species']:
  if val in pattern:
    lis.append('matched')
  else:
    lis.append("notmatched")

iris['new_column'] = lis

Pandas solution is possible by numpy.where and Series.isin :

iris['new_column'] = np.where(iris['species'].isin(pattern), 'matched', 'notmatched')        

Answer 2

I ended up finding an answer through a few different threads.

Here's how I did it

iris = sns.load_dataset('iris')
iris.head()
print (iris.species.value_counts())

pattern = ['setosa', 'virginica']
iris['new_column'] = 0

for index, row in iris.iterrows():
  match = re.match('|'.join(pattern), row.species)
  if match:
    iris.loc[index, "new_column"] = match.group(0)
  else:
    iris.loc[index, "new_column"] = 'no match'


print (iris.new_column.value_counts())

I imagine there's a more efficient way to do this and I also have to specify the column, which isn't ideal. Feel free to comment!