Feature proposal: small auto inferencing edit

[aborruso]

Hi,
if I write

echo 'a=01,b=aaa,c=12.3' | mlr --ojson cat

I have { "a": 01, "b": "aaa", "c": 12.3 } and not { "a": "01", "b": "aaa", "c": 12.3 }. a is a string.

It's a zero padding problem.

What about to add in the auto inferencing process a regex control (if value =~ "^0[0-9]+") that maps correctly this very widespread kind of field value?

I know, I can use -S. And this is another example

echo 'a=00009,b=1.23456' | mlr  --opprint --barred put -S '$out1=$a;$out2=$b'

to have

+-------+---------+-------+---------+
| a     | b       | out1  | out2    |
+-------+---------+-------+---------+
| 00009 | 1.23456 | 00009 | 1.23456 |
+-------+---------+-------+---------+

But -S is global and if I run

echo 'a=00009,b=1.23456' | mlr  --opprint --barred put -S '$out1=$a;$out2=$b*3'

I have error

+-------+---------+-------+---------+
| a     | b       | out1  | out2    |
+-------+---------+-------+---------+
| 00009 | 1.23456 | 00009 | (error) |
+-------+---------+-------+---------+

Thank you

[johnkerl]

@aborruso I believe this is resolved by #756 / #788:

kerl@weryx[][miller]$ echo 'a=00009,b=1.23456' | mlr  --ojson put '$out1=$a;$out2=$b'
{
  "a": "00009",
  "b": 1.23456,
  "out1": "00009",
  "out2": 1.23456
}
kerl@weryx[][miller]$ echo 'a=00009,b=1.23456' | mlr --opprint --barred put '$out1=$a;$out2=$b'
+-------+---------+-------+---------+
| a     | b       | out1  | out2    |
+-------+---------+-------+---------+
| 00009 | 1.23456 | 00009 | 1.23456 |
+-------+---------+-------+---------+

If I'm mistaken please feel free to re-open -- thanks!