DataFrame.tensor

Return a Tensorflow tensor representation of the DataFrame. Only the values in the DataFrame will be returned, the axes labels will be removed.

danfo.DataFrame.tensor [source]

Returns:

return tf.tensor

Note: Tensorflow tensors have single dtype (mainly float32), and will replace any string value with NaN. Use with care.

Examples

Node
Browser
Node
const dfd = require("danfojs-node")
let data = { "A": [-20, 30, 47.3, -20] ,
"B": [34, -4, 5, 6] ,
"C": [20, 20, 30, 30]}
let df = new dfd.DataFrame(data)
let tf_tensor = df.tensor
console.log(tf_tensor.dtype);
console.log(tf_tensor);
tf_tensor.print()
Browser
Output
Output
float32
Tensor {
kept: false,
isDisposedInternal: false,
shape: [ 3, 3 ],
dtype: 'float32',
size: 9,
strides: [ 3 ],
dataId: {},
id: 0,
rankType: '2'
}
Tensor
[[-20 , 34, 20],
[30 , -4, 2 ],
[47.2999992, 5 , 30]]

String values in a Tensor are represented as NaN, so ensure to transform them before working with tensor representations.

Node
Browser
Node
const dfd = require("danfojs-node")
let data = { "Abs": [20.2, 30, 47.3] ,
"Count": [34, 4, 5, 6] ,
"country code": ["NG", "FR", "GH"] }
let df = new dfd.DataFrame(data)
let tf_tensor = df.tensor
console.log(tf_tensor.dtype);
console.log(tf_tensor);
tf_tensor.print()
Browser
Output
Output
float32
Tensor {
kept: false,
isDisposedInternal: false,
shape: [ 3, 3 ],
dtype: 'float32',
size: 9,
strides: [ 3 ],
dataId: {},
id: 0,
rankType: '2'
}
Tensor
[[20.2000008, 34, NaN],
[30 , 4 , NaN],
[47.2999992, 5 , NaN]]