meto - 8 months ago 76

Python Question

Is there a way to create a

`pd.Series`

My code looks like the following, but I'm sure there is a better way:

`import numpy as np`

import pandas as pd

g = ((n, s) for n, s in [("A", 1), ("B", 2), ("C", 3), ("D", 4), ("E", 5)])

arr = np.array(list(g))

ind, val = arr[:, 0], arr[:, 1]

pd.Series(val, index=ind)

Answer

Here's an alternative using the `DataFrame`

constructor:

```
>>> g = ((n, s) for n, s in [("A", 1), ("B", 2), ("C", 3), ("D", 4), ("E", 5)])
>>> pd.DataFrame(g).set_index(0)[1]
0
A 1
B 2
C 3
D 4
E 5
Name: 1, dtype: int64
```

After the DataFrame is constructed, we set the index column and return a Series by selecting column 1.

This avoids the need for any temporary lists so might be more efficient (I haven't tested it yet). It also uses appropriate dtypes for each of the columns (int64 in this case) so it avoids object arrays being created first.

Source (Stackoverflow)