Regex capture groups and use OR statement

Question

I'm trying to create a regex expression that has has multiple conditions separated by | (OR). I want to use capture groups but I'm not getting it to work fully.

3 sample strings:

--- {source-charset: '', encoding-error-limit: '', class: stat-direct, directory: \\\myserver\C\FOLDER\SUB_FOLDER}  
--- {odbc-connect-string-extras: '', server: hello.sample.com, dbname: X_DB, port: '80', class: hello, username: USERX}  
--- {cleaning: 'no', filename: //myserver/D/FOLDER/SUB_FOLDER/File name.xlsx, dataRefreshTime: '', interpretationMode: '0'}

For each sample string I would like the regex to return:

\\\myserver\C\FOLDER\SUB_FOLDER  
X_DB  
//myserver/D/FOLDER/SUB_FOLDER/File name.xlsx

Basically the value after either directory:, dbname: or filename: and ending with } for one of them and , for two.

I've managed to use OR statements to get the three conditions in. regex extract

'directory: [^}]+|dbname: [^,]+|filename: [^,]+'

That returns:

directory: \\\myserver\C\FOLDER\SUB_FOLDER}  
dbname: X__DB,  
filename: //myserver/D/FOLDER/SUB_FOLDER/File name.xlsx,

If I introduce capturing groups I only get the right return for one of the parts: 'directory: ([^}]+)|dbname: ([^,]+)|filename: ([^,]+)'

That returns:

\\\myserver\C\FOLDER\SUB_FOLDER  
null  
null  

I've managed to get it working with a nested regex that takes the result from

'directory: [^}]+|dbname: [^,]+|filename: [^,]+'   

and uses:

': ([^,}]+)'

That gives me the result I want but I would like to do this as one regex. Any help would be greatly appreciated.

/Aron

Answer 1

You could use a negated character class to match not a {, } or a comma, match any of the options in a non capturing group and use and a single capturing group the capture the values:

{[^{]+(?:filename|directory|dbname): ([^,}]+)[^}]*}

Explanation

• { Match {
• [^{]+ Match 1+ times not { using a negated character class
• (?:filename|directory|dbname): Match any of the listed options followed by : and a space
• ( Capture group1
  • [^,}]+ Match 1+ times not , or }
• ) Close group 1
• [^}]*} Match 0+ times not }, then match }

Regex demo