Need help writing a regular expression

Question

These are some samples of the data I am working with(I made some comments on the side):

TSG MUM \n\nBS06-312
RQWE. FKB \n\nBS06-204
NM. JAK \n\nBS06-E05
DB. FKB \n\nBS06-312
IGT. resetk \n\nWender.   //--> special CASE 
ENG I. \n\nWEHN BS06-E06 \n\nENG II \n\nFLEM BS06-203 //--> special CASE: 2 Subjects
ITSI. MUM \n\nBS06-E02
PQT. RIE \n\nBS11-QCR PQT \n\nMARK BS11-QCR \n\nPQT FIS \n\nBS11-QCR //--> special CASE: several Subjects
INC FEY \n\nBS06-309
FU MAT \n\nSKU BS06-309
ABS. DOE \n\nBS06 ABS \n\nVOG BS06 \n\nABS HEI \n\nBS06 ABS \n\nMOR BS06 \n\nABS REM \n\nBS06 ABS \n\nDEI BS06 \n\nABS THA \n\nBS06
ENG III. \n\nGLIT BS06-209 \n\nENG II \n\nWANN BS06-208

These are subjects in a class schedule. The first letters represent the taught subject. After that its the teachers initials separated by a space. The last position is the building and room number.

Sometimes there are several subjects being taught on a specific time.

The data comes from an ics calendar file and I simply copied it here. The new line characters also need to be considered.

I need to extract the subject name, teachers initials and the room number so i can work with it. Any ideas on how to proceed? A complete regex pattern would be ideal.

I am working with php.

Thank you for your help.

Why is there sometimes a dot after the subject, and sometimes not ? — zessx, Jul 08 '14 at 07:55
I dont know about the dot. Its just there sometimes. This is what i did so far: $string = trim(preg_replace('/\s\s+/', ' ', $summary)); $pieces = explode(" ", $string); // its inconsistent — sainjori, Jul 08 '14 at 08:11
It's indeed specially difficult to obtain the good result for each cases. Can you explain with more details the first special case? — Casimir et Hippolyte, Jul 08 '14 at 08:22
The different pieces of information are seperated consistently with a space. The new line is just formatting. — sainjori, Jul 08 '14 at 08:24

score 2 · Accepted Answer · edited May 23 '17 at 12:20

I will exclude this line :

FU MAT \n\nSKU BS06-309

What have we got here ?

FU : subject
MAT : teacher
SKU : ???
BS06-309 : room

Solution :

Anyway, for the rest of the block you can user this regex :

(?:\s|\\n\\n)*(?<subject>\S+(?:\s[IVX]+\.?)?)(?:\s|\\n\\n)+(?<teacher>\S+)(?:\s|\\n\\n)+(?<room>\S+)(?:\s|\\n\\n)*

Regular expression visualization

Details :

(?:\s|\\n\\n)*                   # spaces or \n\n - not caught
(?<subject>\S+(?:\s[IVX]+\.?)?)  # non-spaces plus I., II., III, IV... -> subject
(?:\s|\\n\\n)+                   # spaces or \n\n - not caught
(?<teacher>\S+)                  # non-spaces -> teacher
(?:\s|\\n\\n)+                   # spaces or \n\n - not caught
(?<room>\S+)                     # non-spaces -> room
(?:\s|\\n\\n)*                   # spaces or \n\n - not caught

Result :

+-------+----------+---------+----------+
| MATCH | SUBJECT  | TEACHER | ROOM     |
+-------+----------+---------+----------+
| 1     | TSG      | MUM     | BS06-312 |
| 2     | RQWE.    | FKB     | BS06-204 |
| 3     | NM.      | JAK     | BS06-E05 |
| 4     | DB.      | FKB     | BS06-312 |
| 5     | IGT.     | resetk  | Wender.  |
| 6     | ENG I.   | WEHN    | BS06-E06 |
| 7     | ENG II   | FLEM    | BS06-203 |
| 8     | ITSI.    | MUM     | BS06-E02 |
| 9     | PQT.     | RIE     | BS11-QCR |
| 10    | PQT      | MARK    | BS11-QCR |
| 11    | PQT      | FIS     | BS11-QCR |
| 12    | INC      | FEY     | BS06-309 |
| 13    | ABS.     | DOE     | BS06     |
| 14    | ABS      | VOG     | BS06     |
| 15    | ABS      | HEI     | BS06     |
| 16    | ABS      | MOR     | BS06     |
| 17    | ABS      | REM     | BS06     |
| 18    | ABS      | DEI     | BS06     |
| 19    | ABS      | THA     | BS06     |
| 20    | ENG III. | GLIT    | BS06-209 |
| 21    | ENG II   | WANN    | BS06-208 |
+-------+----------+---------+----------+

Try it :

Demo

Improve it !

There're roman numbers sometimes : ENG I., ENG II...
I assume you'll only numbers from 1 to 39, that's why I only use [IVX]. You can improve this part adding L, C, M... Or using a real regex for roman numbers.

Need help writing a regular expression

1 Answers1