Regex to Select only Pattern Match Files

Question

Ftp server has following files created on daily basis.

FGI_WTYUIO_D_2016_04_16_BS.daily.gzip - BS File
FGI_WTYUIO_D_2016_04_16_BV.daily.gzip - BV File
FGI_GHJK_D_2016_04_16_SATB3.daily.gzip - B3 File
FKI_GHJK_D_2016_04_16_SAT.daily.gzip - BV File
FKI_GHJK_D_2016_04_16_SATB3.daily.gzip - B3 File
FKI_GHJK_D_2016_04_16_SATBS.daily.gzip - BS File
FKI_GHJK_D_2016_04_16_SSD.daily.gzip - Need to Ignore
FKI_GHJK_D_2016_04_16_SSDBS.daily.gzip - Need to Ignore

So, basically there two filetypes

FGI
FKI

and Three Report code for each Filetypes

BS
BV
B3

I need to ignore rest of the files. (SSD files).

I need to write regex pattern inside Javascript to fetch these files. which has following variables.

fileDate - Date ex. 2016_04_16
matchReportCode - ex. BV,BS,B3

So, if fileDate = 2016_04_15 and matchReportCode='SV' (BS,BV). Then I should only fetch following files.

FGI_WTYUIO_D_2016_04_15_BS.daily.gzip - FGI BS File
FGI_WTYUIO_D_2016_04_15_BV.daily.gzip - FGI BV File
FKI_GHJK_D_2016_04_16_SAT.daily.gzip - FKI BV File
FKI_GHJK_D_2016_04_16_SATBS.daily.gzip - FKI BS File

So, if fileDate = 2016_04_19 and matchReportCode='3S' (B3,BS). Then I should only fetch following files.

FGI_WTYUIO_D_2016_04_15_BS.daily.gzip - FGI BS File
FGI_GHJK_D_2016_04_16_SATB3.daily.gzip - FGI B3 File
FKI_GHJK_D_2016_04_16_SATB3.daily.gzip - FKI B3 File
FKI_GHJK_D_2016_04_16_SATBS.daily.gzip - FKI BS File

So far I could only come up with this.

FileRegex = "F[KG]I_.*_D_" + fileDate + "_[A-z]{0,3}L{0,1}[" + matchReportCode + "]{0,1}.daily.gzip";

Can someone please help ? I am new to regex. Thanks.

Correct Regex : FileRegex = "F[KG]I_.*_D_" + fileDate + "_[A-z]{0,3}B{0,1}[" + matchReportCode + "]{0,1}.daily.gzip"; — jigarshah, May 02 '16 at 03:27
[`[A-z]` matches more than you think](http://stackoverflow.com/questions/29771901/why-is-this-regex-allowing-a-caret/29771926#29771926). — Wiktor Stribiżew, May 02 '16 at 12:50

Emma Talbert · Answer 1 · 2016-05-02T19:36:22.753

The following may be a bit better:

FileRegex = "F[KG]I_[^_]+_D_" + fileDate + "_(?!SSD)[a-zA-Z]{0,3}((B[" + matchReportCode + "])|(?<^FKI.*)).daily.gzip";

This will match FKI and FGI file names that have the chosen fileDate and up to three letters preceding the chosen reportCode.

The other changes include changing [A-z], to [a-zA-Z], this is because regex character class range expression uses the ascii representations, and there are characters ([- etc.) between A and z that are not alphabetical (which appears to be your intent).

Also .*_ became [^_]+_, this requires that there be at least one character besides the underscore, prevents the engine from having to backtrack as much, as well as making the regex easier to edit if another segment is added.

I also added a negative lookahead (?!SSD) at the start of the last segment, which requires that the segment not start with SSD.

The or condition at the end ((B[" + matchReportCode + "])|(?<^FKI.*)) requires that either the file match the report code, or that the file name start with FKI (followed by any number of characters to get back to the end). The ^ is the start of line anchor when used outside of a character class ([...]).

How to ignore SSD fles ? Regex will copy that as well , right ? Also, I don't have BV code in FKI type - FKI_GHJK_D_2016_04_16_SAT.daily.gzip - — jigarshah, May 02 '16 at 04:03
See my updated answer. I wasn't sure if the missing report code was a typo, and I can't comment on other posts to clarify because I only have like 10 rep at the moment. — Emma Talbert, May 02 '16 at 04:37

score 0 · Answer 2 · answered May 02 '16 at 02:59

You can use negative lookahead:

var fileDate = '2016_04_19';
var matchReportCode = 'BS';
var re = new RegExp('F[KG]I_\\w+_D_' + fileDate + 
  '_(?!SSD)[\\w\\d]*' + matchReportCode + '?\\.daily\\.gzip');

// re.test('FKI_GHJK_D_2016_04_19_SSDBS.daily.gzip') === false
// re.test('FKI_GHJK_D_2016_04_19_SATBS.daily.gzip') === true

Regex to Select only Pattern Match Files

2 Answers2