string - Multilevel parsing using shell command -

i have file in following format

///// name 1 start_occurrence: occurrence 1 occurrence 2  /// name 2 start_occurance: occurrence 1 occurrence 2  /// name 3 start_occurrence: occurrence 1 occurrence 2 occurrence 3 

all need make count of number of occurrences each name , save them in csv file. can using combination of shell commands? yes can programmatically, looking bunch of shell commands in pipe lined fashion.

"names" can anything. names not come pattern. catch line after /// name. occurrence not have number it, anyline starts occurrence or have occurrence subject of interest.

awk 'c=="thisisname"{b=$0;c="";}$1=="///"{c="thisisname"}$0~/\<occurrence\>/{a[b]+=1;}end{for (i in a){print i" "a[i]}}'  your_file_here 

explain:

if match name start condition ($1=="///"), mark c thisisname.  if name line (c=="thisisname"), mark name line b, , mark c name part ended(c="").  if match occurrence condition ($0~/\<occurrence\>/), make a[b] += 1.  use map remark occurrence time of each name. 

awk use eres, $0~/eres/ means $0 match regex. '\<' , '>' means '\b' in pres