Keep/Drop wildcard

[NicoLaval]

Hi,

Currently, keep & drop only accept componentID.

For Dataset having a large number of columns, when you transpose a dataset building 0/1 filled columns, named var_1, ..., var_N and others columns, it would be nice to be able to define:

ds := ds[keep "var_*", other_var_name, "*_other_pattern"]
//or
ds := ds[drop "var_*"]

What do you think?

[vpinna80]

In order to avoid string interpretation, i would use regexps, though I'm fine with the proposal in general.

Perhaps regexps could be introduced as a new lexer mode, with the slash: https://github.com/antlr/antlr4/blob/master/doc/lexer-rules.md#lexical-modes

There are lexers for regexp that could be adopted without much effort, this is one:
https://github.com/antlr/grammars-v4/blob/c82c128d980f4ce46fb3536f87b06b45b9619922/xsd-regex/regexLexer.g4

Regexp literals could also be added as a new VTL type (scalar or string subset) and used in VTL string functions (instr, replace, match_characters)