REGEX to match 3 or 4 columns of data, one of which might contain spaces

user_1430420 0
Viewed 83

I'm trying to parse, in PHP, a data file describing network nodes. The data file consists of 4 columns, with the 4th column being optional:

<Node name>  <Operating System> <Description> <SSL info>

e.g.

    SIG.SND.SERV                   ZOS        A Big client (MF  LPAR   PROD)
    ADC1                           ZOS        AMEX                           SSL
    ADEPTRA.GB1.PROD               LINUX      BANK OF SOMEWHERE            TLS
    ADEPTRA.GB2.CQA                LINUX      BANK OF SOMEWHERE            TLS
    AIX.EG3C                       UNIX       BARCLAYS                       S+
    AIX.EG3P                       UNIX       BARCLAYS                       S+
    ADEPTRA.GB2.CQA                LINUX      FICO (ADEPTRA) LTD            TLS
    AIX.RMWDEV1                    UNIX       FDCS
    AIX.RMWPROD1                   UNIX       FDCS
    AIX-EFXWRWCK01                 UNIX       EQUIFAX EUROPE
    ANB-DRC-CDFDI                  NT         ARAB NATIONAL BANK (ANB)      SSL

My problems lie with column3, the description, which may have spaces and column 4 with may be absent.

I've tried the following REGEXes (with the REGEX checker at https://regex101.com/ :

(\S+)\s+(\S+)\s+(\S.+?\S)\s+(\S.+)?

(This was my attempt to express 'non-whitespace followed by some spaces, non-whitespace followed by some spaces, characters delimited by non-whitespace followed by at least 2 spaces

which works except when the 4th column is absent. My attempts to make the last spaces and characters optional results in columns 3 and 4 matching together:

(\S+)\s{2,} (\S+)\s{2,} (\S.+\S)(\s{2,})? ?(\S.+)?

I also tried to create a REGEX that specified 'Any characters deliminated by non-whitespace followed by either at least 2 spaces or the end of the string' :

(\S.+?\S)[\s{2}|$]

which is also 'close but no cigar'.

Sadly yes - that description in the first line does have multiple spaces between the words.

Added a link to one of my attempts: https://regex101.com/r/k9JlZu/1

php regex
2 Answers

You may use this regex to capture 3 or 4 groups in each line:

^(\S+)\h+(\S+)\h+(\w+(?:\h\w+)*(?:\h*\([^)]+\)(?:\h\w+)?)?)(?:\h{2,}(\S+))?$

RegEx Demo

RegEx Details:

  • ^: Start
  • (\S+): Match 1+ non-whitespace characters in capture group #1
  • \h+: Match 1+ whitespaces
  • (\S+): Match 1+ non-whitespace characters in capture group #2
  • \h+: Match 1+ whitespaces
  • (\w+(?:\h\w+)*(?:\h*\([^)]+\))?): Capture group #3 to match spaces separated words optionally followed by a (...) string
  • (?:\h+(\S+))?: An optional non-capture group in the end to capture 1+ non-whitespace characters in capture group #4
  • $: End
user_5482250

Code: (Demo)

preg_match_all(
    '/^(\S+)\h+(\S+)\h+(\S+(?:\h\S+|\h+\([^)]+\))*)(?:\h+(\S+))?/m',
    $txt,
    $m
);
var_export($m);

It only seems that the 3rd group is the tricky one.

I recommend a shorter subpattern than what anubhava is currently showing:

(                 # capture group #3
  \S+             # one or more visible characters
  (?:             # a non-capturing group
    \h\S+         # a single horizontal space followed by one or more visible characters
    |             # or
    \h+\([^)]+\)  # one or more horizontal spaces, opening parenthesis, one or more non-closing parentheses, closing parenthesis
  )*              # zero or more times
)

The fourth/optional component is caputured inside of a non-capturing group with a zero or one quantifier (?).

Pattern comparisons: (anubhava (423 steps) vs mickmackua (396 steps))

// anubhava
/^(\S+)\h+(\S+)\h+(\w+(?:\h\w+)*(?:\h*\([^)]+\)(?:\h\w+)?)?)(?:\h{2,}(\S+))?$/m

// mickmackusa
/^(\S+)\h+(\S+)\h+(\S+(?:\h\S+|\h+\([^)]+\))*)(?:\h+(\S+))?/m

So my pattern is both more concise/D.R.Y. and more efficient in terms of steps.

user_29434030
Related