Data conversion error from csv to DataPreview mode

[gaille34]

Hi,
When I convert or visualize a .csv file with a value XX.0 in the first line of a column all values of the others lines of the same column loose the number after the point: -41.6 became -41.
Please see the 2 attached screenshot.
Best regards

[IlyaOrson]

I hit a similar issue with this data.zip:

t,x1,x2,c1
0.0,1.0,3.0,1.453712e-11
0.01,0.99932516,3.0006747,1.4536232e-11
0.02,0.99864984,3.0013502,1.4535234e-11
0.03,0.99797404,3.002026,1.4534291e-11
...

Line plots are broken for the x1 and x2 columns, the other ones are parsed correctly.

[RandomFractals]

that's because your first line has ints and data type is detected based on the first data line.

also, not an error. I'd call it incorrect data precision issue at best :)

[gaille34]

Ok I see better now.
Could it be possible to configure/manage this "precision" manually ?
Today I have to be cautious on large datasets and I need to put 43.001

to have this result:

Best regards

[RandomFractals]

yeah, I need to see if I can include more data rows for detecting ints and decimals, and there is a separate ticket for custom data types that I plan to address at some point (#156)

[Meussdorffer]

that's because your first line has ints and data type is detected based on the first data line.

also, not an error. I'd call it incorrect data precision issue at best :)

Not an error in the code, but certainly an error in methodology. I'd suggest implementing some random sampling to infer type / precision instead of naively using the first row. This tool is unusable for me because of this bug.

[Meussdorffer]

@Meussdorffer ok! I am always open to suggestions from devs with 0 commits :)

Always a great look when you patronize your users for making legitimate suggestions.

[rfrenchseti]

I came here to report this exact same thing. I was demoing this extension and my CSV file starts with:

angle,x,y
0.0,1.0,0.0
0.06283185307179587,0.9980267284282716,0.06279051952931337
0.12566370614359174,0.9921147013144779,0.12533323356430426
0.1884955592153876,0.9822872507286887,0.18738131458572463

Note the first line does have a floating point indication, it just happens to be an integer value. The whole column is marked as integer. This is definitely a problem. Can you infer the data type by looking at the format of the number instead of the actual value?