Add comma separated count from list of string

I have a column named diff_2 in my df, which is of this form:

diff_2

error at /Users, in API GET /projects/{projectId} the response property 'id' became optional for the status '200' [response-property-became-optional]. 

But what I want to achieve is a count for both type of changes, comma separated.

I am not sure how this can be done, any suggestions or ideas would be really helpful.

Here is an updated answer which includes combining the count results and adding them back into the original data frame.

I have devided the solutiion into three stages for clarity:

1. Count the different error messages in content
2. Merge the error messages and counts into comma seperated strings
3. Add the results back into the original dataframe as 2 columns with constant values.

Stage 1: Do the counting.

# Count the number of different content items.
cnt = df.groupby('content').count()

# My example only has the `diff_2` and `content` coluumns. 
# If your daaframe has additional columns they should be stripped like this:
cnt = cnt.loc[:,['diff_2']].copy()

# Label the count column appropriately.
cnt.columns = ['count']

# Move the `content` column from the index to a column.
cnt.reset_index(inplace=True)

# Convert the columns to the string data type so that they can be combined.
cnt = cnt.astype(str)
print('Stage 1')
print(cnt)

The counting results look like this:

Stage 2: Merge the results as comma delimited strings.

# Make a simple function that takes a series of strings and combines them with a comma seperator.
def combine(srs): return srs.str.cat(sep=',')

# Apply this function to both columns of the count dataframe.
combined = cnt.apply(combine)
print()
print('Stage 2')
print(combined)

• The second stage result is a Series, where each row is a string complilation of a columns in cnt.
• The index of the series is the names of the columns in cnt.

The results look like this:

• The stage 2 results should be the same for every row in the original df.
• Each row in combined can be added to the original df as columns with constants:

Stage 3: Add the results back into the original DataFrame.

for label in combined.index:
    # Add a prefix to the column name to avoid duplicate names
    col_name = 'Merged_' + label
    # Set the value as a column in df with constant values for each row.
    df[col_name] = combined.at[label]
print()
print('The final stage:\t Merging the reults with the original df')
print(df.info())

The result is two new columns in df:

...