CODEWITHSUNDEEP
pythonpython-3.xpandasdtype
Sh2021
Sh2021

Reputation: 11

read_csv specifying dtype of a particular column to read as list

I am trying to read a CSV file with the read_csv method. The column item_number needs to be read as a list.

Example : 'item_number' : ['123456']

df = pd.read_csv(filepath, sep =',', dtype = {'item_number' = list})

dtype = {'item_number' = list} ->does not work, while when I write it as dtype = {'item_number' = str} , it does work. Tried using converters but want to do it with dtype.

Upvotes: 1

Views: 1253

Answers (2)

U13-Forward
U13-Forward

Reputation: 71560

I think you mean by?

df = pd.read_csv(filepath, sep =',', dtypes = {'item_number': list})

But I am not sure with your dataset if that would work or not, so I guess you need:

df = pd.read_csv(filepath, sep =',', dtypes = {'item_number': pd.eval})

Edit:

Try:

df = pd.read_csv(filepath, sep =',')
df['item_number'] = df['item_number'].str[1:-1].map(list)

Upvotes: 1

Mortz
Mortz

Reputation: 4879

I don't think list is a valid pandas dtype. You can read it in as a string and subsequently convert it to a list

file_str = '''1|[123456]
2|[345678]'''

df = pd.read_csv(io.StringIO(file_str), sep='|', header=None)
df[2] = df[1].apply(lambda x: list(x[1:-1]))
print(df)
#   0         1                   2
#0  1  [123456]  [1, 2, 3, 4, 5, 6]
#1  2  [345678]  [3, 4, 5, 6, 7, 8]

Upvotes: 1

Related Questions