正则表达式从复杂字符串中提取子字符串

Question

I've a string that can have in two formats,

First Format:

route-policy testPolicy
  if (destination in pfx_manju) then
    set extcommunity rt (10429:103) additive
  endif
end-policy

Second Format:

route-policy testPolicy
  if (destination in EXP1) then
    set extcommunity rt (27699:352002, 2.2.2.2:98) additive
  elseif (destination in pfx_manju) then
    set extcommunity rt (27699:339600, 27679:339700, 1.1.1.1:6763, 65536:45633) additive
  elseif (destination in EXP5) then
    drop
  endif
end-policy

Third Format:

route-policy EXPORTRP1
  if (destination in EXP1) or (destination in EXP2) then
    set extcommunity rt (27699:352002, 2.2.2.2:98) additive
  elseif (destination in pfx_manju) or (destination in EXP4) then
    set extcommunity rt (27699:339600, 27679:339700, 1.1.1.1:6763, 65536:45633) additive
  elseif (destination in EXP5) or (destination in EXP6) then
    drop
  endif
end-policy

So here the complete text is coming as string. The string either single condition (if) or multiple conditions (elseIf conditions).

From the above string I want to extract the rt values for one hard coded policy (pfx_manju). I can extract the rt values with below regex,

final String regex = "rt \\(([^)]+)\\)";

Now the problem is, I want to extract the sub string which is belongs to the hard coded policy (pfx_manju).

So condition is get a sub string with starting position is index of pfx_manju and ending position is the subsequent endif or elseif .

So I want the sub sting output for above examples as mentioned below,

First Sub String:

  pfx_manju) then
    set extcommunity rt (10429:103) additive

Second Sub String:

pfx_manju) then
        set extcommunity rt (27699:339600, 27679:339700, 1.1.1.1:6763, 65536:45633) additive

Third Sub String:

pfx_manju) or (destination in EXP4) then
        set extcommunity rt (27699:339600, 27679:339700, 1.1.1.1:6763, 65536:45633) additive

Any suggestions with optimistic solution will be appreciated

Answer 1

The following regex should work

pfx_manju\)[\s\S]*?rt \(([^)]+)\) additive

The regex matches string starting from pfx_manju condition all the way to rt values, that means it captures the rt values when there is only pfx_manju condition. if you are using java, you need to scape, \ .

See regex demo