ENH: add new method `DataFrame.repeat`

[Zeroto521]

Is your feature request related to a problem?

I sometimes would generate some faker but having rule data.
Just like in a (2d) data, repeat some rows or columns serial times.

The steps I do:

1. This is a DataFrame called df.
2. value = df.values # change DataFrame to ndarray
3. use np.repeat or np.tile method to repeat.

Describe the solution you'd like

I thought why not add a method repeat to DataFrame, since Series has already owned the repeat method.

About the DataFrame.repeat parameters

Similar to numpy.repeat(a, repeats, axis=None)

def repeat(
    self,
    repeats: int | list[int],
    axis: int | str = 0,
) -> pd.DataFrame | None:
    ...

DataFrame is a 2d data structure, so it is hard to repeat each element.
And the aim of DataFrame.repeat is to repeat one row/column or some rows/columns.

So the axis argument must be specific to 0 or 1.

And if the repeats is a single value int, that means repeat all row/column repeats times.

Examples of this method

>>> df = pd.DataFrame({'a': [1, 2], 'b':[3, 4]})
>>> df
    a  b
0  1  3
1  2  4

Each row repeat two times.

>>> df.repeat(2)
    a  b
0  1  3
0  1  3
1  2  4
1  2  4

Each column repeat two times.

>>> df.repeat(2, 1)
    a  a  b  b
0  1  1  3  3
1  2  2  4  4

``a`` column repeat 1 times, ``b`` column repeat 2 times.

>>> df.repeat([1, 2], 1)
    a  b  b
0  1  3  3
1  2  4  4

[samukweku]

Why not repeat the index and reindex the data frame with that

[Zeroto521]

Why not repeat the index and reindex the data frame with that

To more think about repeat to index and repeat.
'reindex' doesn't only reset index, but also reset column.
To number type, could be reset from 0 to n.
To other types (ex. string) could be reset to what type?

>>> df = pd.DataFrame({'a': [1, 2], 'b':[3, 4]}, index=[2, 3])
>>> df
    a  b
2  1  3
3  2  4

# Repeat row 2 times without resetting the index.

>>> df.repeat(2)
    a  b
2  1  3
2  1  3
3  2  4
3  2  4

# Repeat row 2 times with resetting the index.
# Actually, I thought 'repeat and reindex' is a little bit weird.
# There should notice the index that is not starting from 0.
# But it's ok to number index, it still could be reset

>>>  df.repeat(2)
    a  b
0  1  3
1  1  3
2  2  4
3  2  4

# repeat column 2 times with resetting
>>> df.repeat(2, 1)
# What's the column?
???

Series.repeat's examples are the following.

>>> s = pd.Series(['a', 'b', 'c'])
>>> s
0    a
1    b
2    c
dtype: object
>>> s.repeat(2)
0    a
0    a
1    b
1    b
2    c
2    c
dtype: object
>>> s.repeat([1, 2, 3])
0    a
1    b
1    b
2    c
2    c
2    c
dtype: object

[samukweku]

I'm not sure if it makes sense to have duplicate columns, but this is how you can repeat a dataframe with the index(or columns, which is an index too):

df.reindex(df.index.repeat(2))  # repeat on the index

   a  b
0  1  3
0  1  3
1  2  4
1  2  4


df.reindex(columns=df.columns.repeat(2)) # repeat on the columns

   a  a  b  b
0  1  1  3  3
1  2  2  4  4


 df.reindex(columns=df.columns.repeat([1,2])) # repeat column a once, b twice

   a  b  b
0  1  3  3
1  2  4  4

Reindex does not reset the index, it just conforms it to the new index. It only works if the existing index is unique.

[jreback]

-1 on adding even more api to a huge api

happy to take an example though for the doc string

[simonjayhawkins]

-1 on adding even more api to a huge api

agreed.

The numpy array protocol allows the use of the numpy api directly, so no need for pandas to add.

so np.repeat(df, 2, axis=1) should work. doesn't currently. raises ValueError: Shape of passed values is (2, 4), indices imply (2, 2)