Using pd.DataFrame(tensor) is abnormally slow, you can make the following modifications

[YeahNew]

• I have checked that this issue has not already been reported.

• I have confirmed this bug exists on the latest version of pandas.

• I have confirmed this bug exists on the master branch of pandas.

Reproducible Example

import numpy as np
import pandas as pd
import torch

row = 700000
col = 64
val_numpy = np.random.rand(row, col)
val_tensor = torch.randn(row, col)

numpy_pd_start_time = time.time()
va_numpy_pd = pd.DataFrame(val_numpy)
numpy_pd_end_time = time.time()
print("numpy to pd time:{:.4f}s".
      format(numpy_pd_end_time - numpy_pd_start_time))

tensor_numpy_pd_start_time = time.time()
val_tensor_pd1 = pd.DataFrame(val_tensor.numpy())
tensor_numpy_pd_end_time = time.time()
print("tensor to numpy to pd time:{:.4f} s".
      format(tensor_numpy_pd_end_time - tensor_numpy_pd_start_time))

tensor_pd_start_time = time.time()
val_tensor_pd2 = pd.DataFrame(val_tensor)
tensor_pd_end_time = time.time()
print("tensor to pd time:{:.4f} s".
      format(tensor_pd_end_time - tensor_pd_start_time))

Issue Description

Recently, using pd.DataFrame() to convert data of type torch.tensor to pandas DataFrame is very slow, while converting tensor to numpy and then to pandas DataFrame is very fast. The test code is shown in the Reproducible Example.
The code prints as follows:

numpy to pd time: 0.0013s
tensor to numpy to pd time:0.0005s
tensor to pd time:220.5251s

Then I read the source code and found that if the data accepted by pd.DataFrame() is tensor, tensor will be processed as list_like (line 682 in https://github.com/pandas-dev/pandas/blob/master/pandas/core/ frame.py) .
Mainly time-consuming in the following three stages:

data = list(data)：2.5952s
nested_data_to_arrays: 214.7532s
arrays_to_mgr:2.5987s

In the nested_data_to_arrays stage, a large number of data type conversion operations are involved, the row-list is converted to col-list, and the operation is read by row.This will take a long time.

Sure，This method of use may not be appropriate, but now torch.tensor is widely used, and it is inevitable that it will be used directly in this way, resulting in low efficiency. So can you add a comment at line 467 in frame.py, like this: If data is a torch.tensor, you can transform it to numpy first(tensor.numpy()).
Or can I submit a PR? When it is judged that the input parameter is tensor, execute the conversion, and then execute the ''elif isinstance(data, (np.ndarray, Series, Index))'' judgment.

Looking forward to your reply ~

Installed Versions

pandas.version == 1.3.4

[twoertwein]

I'm not sure how happy people would be adding pytorch as a dependency to pandas. (Could use if hasattr(data, "numpy")?)

Based on the documentation:

one might expect that a pytorch tensor will be treated as an iterable.

It might be worth extending the documentation. Something along the lines of "For best performance, iterable objects, such as a Pytorch Tensor, that can efficiently be converted to a Numpy Array, should be converted before passing it to pd.DataFrame."

[jorisvandenbossche]

I am not super familiar with pytorch, but I suppose they support the array interface? If we don't do that yet, I think we can certainly ensure to treat all objects like that as arrays instead of list-likes.

[YeahNew]

I am not super familiar with pytorch, but I suppose they support the array interface? If we don't do that yet, I think we can certainly ensure to treat all objects like that as arrays instead of list-likes.

Yes, I think they support the array interface, and it is easy to convert between Tensor and Numpy. If the two data types are the same, the memory will be shared after conversion. Judging from the above test results, it is indeed not suitable to convert tensor as list-likes.

[YeahNew]

I'm not sure how happy people would be adding pytorch as a dependency to pandas. (Could use if hasattr(data, "numpy")?)

Based on the documentation: one might expect that a pytorch tensor will be treated as an iterable.

It might be worth extending the documentation. Something along the lines of "For best performance, iterable objects, such as a Pytorch Tensor, that can efficiently be converted to a Numpy Array, should be converted before passing it to pd.DataFrame."

Yes, I think it is appropriate to add such a comment, because it is likely that someone will directly use pd.DataFrame(tensor) to create a DataFrame, which will not report an error, but the performance is very low.
Or Can I submit a PR to modify it?

[attack68]

PR is welcome.

[jbrockmendel]

IIRC from similar issues checking for an __array__ method in sanitize_array was a best-guess for a place to start

[YeahNew]

IIRC from similar issues checking for an __array__ method in sanitize_array was a best-guess for a place to start

Sorry, I don't understand what you mean? Would you like to describe it in detail?