# How to Fix: Data must be 1-dimensional

You might encounter the following error when trying to convert Numpy arrays to a pandas dataframe.

Exception: Data must be 1-dimensional

## 1. Reproduce the Error

# import numpy and pandas
import numpy as np
import pandas as pd

# Create a numpy array of data:
X = np.array([5, 2, 3, 4, 10, 11, 14]).reshape(-1, 1)
Y = np.array([3, 1, 2, 5, 14, 15, 16]).reshape(-1, 1)

# convert to the dataframe
df_1 = pd.DataFrame ({'X':X,'Y':Y}, index=range(1,8))
print (df_1)

Output:

Exception: Data must be 1-dimensional

## 2. Why the Error Happens

It happens because pd.DataFrame is expecting to have 1-D numpy arrays or lists, since it is how columns within a dataframe should be. However, when you use reshape(-1,1), the 1-D array becomes a 2-D array.

We can print out with and without reshape(-1, 1) to see the difference.

X_without = np.array([5, 2, 3, 4, 10, 11, 14])
print('X without reshape(-1,1):\n', X_without )

X_with = np.array([5, 2, 3, 4, 10, 11, 14]).reshape(-1,1)
print('X with reshape(-1,1):\n', X_with)

Output:

X without reshape(-1,1):
[ 5  2  3  4 10 11 14]

X with reshape(-1,1):
[[ 5]
[ 2]
[ 3]
[ 4]
[10]
[11]
[14]]

We can check the dimension and shape of X to illustrate that.

print('X_without dimension is:\n',np.ndim(X_without))
print('X_without shape is: \n',np.shape(X_without))

print('X_with dimension is:\n',np.ndim(X_with))
print('X_with shape is: \n',np.shape(X_with))

Output:

X_without dimension is:
1
X_without shape is:
(7,)

X_with dimension is:
2
X_with shape is:
(7, 1)

## 3. How to Fix the Error

### Method 1: Remove reshape(-1,1)

To fix the error of “Data must be 1-dimensional“, You can remove the reshape(-1,1) to make sure that X and Y are 1-D arrays. The following is the code.

# import numpy and pandas
import numpy as np
import pandas as pd

# Create a numpy array of data:
X = np.array([5, 2, 3, 4, 10, 11, 14])
Y = np.array([3, 1, 2, 5, 14, 15, 16])

# convert to the dataframe
df_1 = pd.DataFrame ({'X':X,'Y':Y}, index=range(1,8))
print (df_1)

Output:

X   Y
1   5   3
2   2   1
3   3   2
4   4   5
5  10  14
6  11  15
7  14  16

## Method 2: use np.ravel()

np.ravel() returns a contiguous flattened array. Thus, we can use that to change 2-D arrays to 1-D arrays.

# import numpy and pandas
import numpy as np
import pandas as pd

# Create a numpy array of data:
X = np.array([5, 2, 3, 4, 10, 11, 14]).reshape(-1, 1)
Y = np.array([3, 1, 2, 5, 14, 15, 16]).reshape(-1, 1)

# convert to the dataframe
df_1 = pd.DataFrame ({'X':X.ravel(),'Y':Y.ravel()}, index=range(1,8))
print (df_1)

Output:

X   Y
1   5   3
2   2   1
3   3   2
4   4   5
5  10  14
6  11  15
7  14  16