Creating a pandas data frame of a specific size

Question

In R, I can do something like this:

myvec <- seq(from =  5, to = 10)^2
mydf <- data.frame(matrix(data = myvec, ncol = 3,byrow = TRUE))
> mydf
  X1 X2  X3
1 25 36  49
2 64 81 100

Notice I can specfiy the shape of the data frame by passing in an ncol parameter. I can then fill it either byrow or bycolumn (in this case by row).

If I were to replicate this in Python/Pandas, it's easy enough to create the sequence:

myData = [x**2 for x in range(5,11) ]

However, how do easily make a dataframe of the same size? I can do something like:

myDF = pd.DataFrame(data = myData)

But what would be the parameters to specify the column/row dimensions?

Answer 1

One way to make a pandas dataframe of the size you wish is to provide index and column values on the creation of the dataframe.

df = pd.DataFrame(index=range(numRows),columns=range(numCols))

This creates a dataframe full of nan's where all columns are of data type object.

Answer 2

Use reshape to specify the number of columns (or rows):

import numpy as np
import pandas as pd

myvec = np.arange(5, 11)**2
mydf = pd.DataFrame(myvec.reshape(-1, 3))

yields

    0   1    2
0  25  36   49
1  64  81  100

---

When calling reshape you are allowed to specify the length of one axis as -1. reshape replaces the -1 with whatever integer makes sense. For example, if myvec.size is 6, and one axis is of length 3, then the other axis has to be of length 6/3 = 2. So the -1 is replaced by 2, and so myvec.reshape(-1, 3) returns an array of shape (2, 3) -- 2 row and 3 columns.