CODEWITHSUNDEEP
pythonpython-3.xpandasdataframe
Souvik Ray
Souvik Ray

Reputation: 3018

How to get a list of strings for a pandas dataframe?

I have a dataframe which I obtain after doing certain operations. This is how my dataframe looks like

        date metric  stat         val       device
0 2018-03-21    cpu  mean   76.724406  10.41.100.1
3 2018-03-21    cpu   std  124.285789  10.41.100.1

Now I want to convert it to a list of strings stored inside a list, something like below

lis = [["2018-03-21", "cpu", "mean", "76.724406", "10.41.100.1"],
["2018-03-21", "cpu", "mean", "124.285789", "10.41.100.1"]]

I did something like this

for i in df:
    print(df[i].tolist())

But I get something like this

[Timestamp('2018-03-21 00:00:00'), Timestamp('2018-03-21 00:00:00')]
['cpu', 'cpu']
['mean', 'std']
[76.72440613174048, 124.28578926665278]
['10.41.100.1', '10.41.100.1']

But I want the format to be like mentioned above. How can I do this?

Upvotes: 1

Views: 3056

Answers (2)

jpp
jpp

Reputation: 164613

The accepted solution by jezrael works.

However, I would suggest you take advantage of the numpy representation directly, since this is a more efficient way of storing and manipulating data.

It is rarely the case that a list is more useful than the numpy representation when dealing with structured data.

You can do this via:

res = df.values.astype(str)

This returns dtype <U11, while df.astype(str).values returns dtype Object. This is a significant difference in how the data is stored internally.

Upvotes: 1

jezrael
jezrael

Reputation: 862396

I think need astype for strings first, then convert DataFrame to numpy array by values and last for list call numpy.ndarray.tolist:

lis = df.astype(str).values.tolist()
print (lis)
[['2018-03-21', 'cpu', 'mean', '76.724406', '10.41.100.1'], 
 ['2018-03-21', 'cpu', 'std', '124.285789', '10.41.100.1']]

Upvotes: 1

Related Questions