Unsigned numpy uint64 unsupported

Hi, When I used a pandas array of uint64 on plotly, the numbers appear negative.
I know that they cannot be as i force the type to unsigned when i append:
tempDataFrame[YaxisIn].astype(np.uint64)
In this call:
traces.append(go.Scattergl(
x = tempDataFrame[XaxisIn],
y = tempDataFrame[YaxisIn].astype(np.uint64),
mode = 'markers',
marker = dict (line = dict(width = 1)),
name = database + "." + TablesIn + ".notUseful"
)
)

[jonmmease]

Hi @suleimanmashini,

Please provide a full example that demonstrates the issue, with data and imports, formatted in a "fenced code block" https://www.google.com/url?sa=t&source=web&rct=j&url=https://help.github.com/articles/creating-and-highlighting-code-blocks/&ved=2ahUKEwjMq--AgJ_dAhWQl-AKHSQrCG4QFjAAegQICBAD&usg=AOvVaw297etkhd2-RtKM-CXyUZ9h.

Also please include the version of plotly.py you are using.

Thanks!

Hi @jonmmease ,
ofcourse Heres a code block that recreates the problem. I'm currently using plotly 3.0.2. I appreciate the help!

import plotly.graph_objs as go
import plotly
import pandas as pd
import numpy as np

print(plotly.__version__)

d = { 'xaxis': [1,2,3,4], 'yaxis': [4173500704,4147505216,4147504160,4143500800]}

df = pd.DataFrame(data=d, dtype=np.uint64)

traces = []

print(df.dtypes)

traces.append(go.Scattergl(
    x = df['xaxis'],
    y = df['yaxis'],
    mode = 'markers',
    marker = dict (line = dict(width = 1)),
    name = "sample data"
    )
)

plotly.offline.plot({
    "data": traces
}, auto_open=True)

This is what I get when running the sample commented above:

[jonmmease]

Hi @suleimanmashini ,

Could you try again with plotly.py 3.2.0 (released yesterday)? When I run your example with this version I don't see this issue:

If that doens't fix it for you, can you post what version of numpy you have installed?

I did notice that trying to plot a figure with a uint64 array breaks the FigureWidget serialization, so I'm going to mark this as a bug either way.

Hey,

updating it fixed the problem for me. Thanks!

[jonmmease]

The FigureWidget issue noted above has been fixed in #1163