Ziru Niu created ARROW-10132:
--------------------------------
Summary: Considers scientific notation when inferring schema from
csv
Key: ARROW-10132
URL: https://issues.apache.org/jira/browse/ARROW-10132
Project: Apache Arrow
Issue Type: Improvement
Components: Rust
Affects Versions: 1.0.1
Environment: Ubuntu
Reporter: Ziru Niu
||col||
|1.2|
|1.3e-2|
|1.4|
Currently this column would be inferred as Utf8 type, since
csv::reader::DECIMAL_RE is defined as r"^-?(\d+\.\d+)$". Maybe we could change
this to r"^-?(\d+\.\d+)(e-?(\d+))?$" or similar stuff to allow scientific
notation of real number inferred as float?
(The RE I currently proposed doesn't handle "5e-4" correctly though)
And I would wish we could infer "3." or ".3" as float too. I will come up with
an exact RE when I get time.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
