Closed kforcodeai closed 2 years ago
I think the new solution can solve your issue -- see example below:
Let's say we have a csv file test.csv
:
Col_A, Col_B
, 1
2, 3
245.0,
And if we read it via:
df = pd.read_csv("test.csv", converters={"Col_A": str})
We have
Test | B |
---|---|
1 | |
2 | 3 |
245.0
(There's no .0
for 2 in the 2nd row and 1st col.
Fixes # https://github.com/Layout-Parser/layout-parser/issues/94#issue-1040387217
94
The issue was, all digit sequences were inferred as float, with this fix all text (numeric + non-numeric) will be inferred as string and the user can change it to their desired data type. But with this fix, the user will be required to change the numeric data type columns. i could not find any better solution other than this.