RegEx: capture entire group content

Question

I am writing a parser for some Oracle commands, like

LOAD DATA
  INFILE  /DD/DATEN
TRUNCATE
PRESERVE BLANKS
INTO TABLE aaa.bbb
( some parameters... )

I already created a regex to match the entire command. I am now looking for a way to capture the name of the input file ("/DD/DATEN" for instance here). My problem is that using the following regex will only return the last character of the first group ("N").

^\s*LOAD DATA\s*INFILE\s*(\w|\\|/)+\s*$

Debuggex Demo

Any ideas? Many thanks in advance

EDIT: following @HamZa 's question, here would be the entire regex to parse Oracle LOAD DATA INFILE command (simplified though):

^\s*LOAD DATA\s*INFILE\s*((?:\w|\\|/)+)\s*((?:TRUNCATE|PRESERVE BLANKS)\s*){0,2}\s*INTO TABLE\s*((?:\w|\.)+)\s*\(\s*((\w+)\s*POSITION\s*\(\s*\d+\s*\:\s*\d+\s*\)\s*((DATE\s*\(\s*(\d+)\s*\)\s*\"YYYY-MM-DD\")|(INTEGER EXTERNAL)|(CHAR\s*\(\s*(\d+)\s*\)))\s*\,{0,1}\s*)+\)\s*$

Debuggex Demo

Answer 1

Let's point out the wrongdoer in your regex (\w|\\|/)+. What happens here ?
You're matching either a word character or a back/forwardslash and putting it in group 1 (\w|\\|/) after that you're telling the regex engine to do this one or more times +. What you actually want is to match those characters several times before grouping them. So you might use a non-matching group (?:) : ((?:\w|\\|/)+).

You might notice that you could just use a character class after all ([\w\\/]+). Hence, your regex could look like

^\s*LOAD DATA\s*INFILE\s*([\w\\/]+)\s*$

On a side note: that end anchor $ will cause your regex to fail if you're not using multiline mode. Or is it that you intentionally didn't post the full regex :) ?

Answer 2

Not tested but...

^\s*LOAD DATA\s*INFILE\s*(\S+)\s*$