Nested regex replacement - regex

Question

I need to create the laravel migrations, so I have converted my SQL script to a laravel migration format using "replacement in files" with regular expressions from Sublime Text.
My problem is that i have to replace in the following string the '#' character by the 'tablename' in about 70 tables:
Schema::table('tablename', function($table) {
$table->dropForeign('#_columnname_foreign');
});
Actually I can do this using the following expression:
(Schema::table\('([a-z]+)',[\s]*function\(\$table\)[\s]*{[\s]*\$table->dropForeign\(')#(_[a-z_]+'\);)
And in the replace field:
$1$2$3
but I don't know how to do when the table has more than one fk:
Schema::table('tablename1', function($table) {
$table->dropForeign('#_field1_foreign');
$table->dropForeign('#_field2_foreign');
$table->dropForeign('#_field3_foreign');
$table->dropForeign('#_field4_foreign');
$table->dropForeign('#_field5_foreign');
$table->dropForeign('#_field6_foreign');
});
I have been using this site to validate my regular expressions RegExr

Answer 1

It is not an easy task for a regex in Sublime Text. The only way to do it with a regex is to make sure you capture the function singature with the optional number of table-dropForeign lines (matched lazily), and replace #s on the next line.
The regex below requires clicking Replace All multiple times until all matches are found.
(Schema::table\('([a-z0-9]+)',\s*function\(\$table\)\s*{(?:\s*\$table->dropForeign\('[a-z0-9]+_\w+'\);)*?\s*\$table->dropForeign\(')#(_\w+'\);)
Replacement is $1$2$3. See this regex demo, where you may replace the # in the second block manually with the table name and see how the match goes further.
Details:
(Schema::table\('([a-z0-9]+)',\s*function\(\$table\)\s*{(?:\s*\$table->dropForeign\('[a-z0-9]+_\w+'\);)*?\s*\$table->dropForeign\(') - Group 1 capturing:
Schema::table\(' - literal Schema::table(' substring
([a-z0-9]+) - Group 2 capturing 1+ alphanumerics (do not check Match Case option to also match uppercase ASCII letters)
',\s* - a comma and 0+ whitespaces
function\(\$table\) - a literal text function($table)
\s* - 0+ whitespaces
{ - a literal { (in SublimeText 2, it requires escaping)
(?:\s*\$table->dropForeign\('[a-z0-9]+_\w+'\);)*? - 0+ sequences, but as few as possible, matching:
\s*\$table->dropForeign\(' - 0+ whitespaces and then a literal text `$table->dropForeign('
[a-z0-9]+_\w+ - 1+ alphanumerics, _ and 1+ digits, letters or underscores (\w+)
'\); - a literal substring ');
\s* - 0+ whitespaces
\$table->dropForeign\(' - a literal text $table->dropForeign('
# - a matched # symbol to be replaced
(_\w+'\);) - Group 2 capturing:
_ - an underscore
\w+ - 1 or more letters, digits or underscores
'\); - a literal substring ');
NOTE: The issue I thought I found was related to an unescaped { that causes a regex failure in Sublime Text 2. In Sublime Text 3, the { in the regex does not have to be escaped.