如何从多行标准输入中仅提取特定的字符串匹配行

Question

I have a standard input piped from another bash function output in the sample form:

fruit=apple
dog=doberman
car=bmw
flower=tulip
---
fruit=orange
dog=corgy
car=ford
flower=rose
---

...and so on repeated thousands of times...

From this multiple repeating piped standard output I want to extract only the values matching say fruit and car, like so:

apple
bmw
---
orange
ford
---

It would be even better if the sed/awk magic can output those in csv table-like format, like so:

apple,bmw
orange,ford

I have only basic knowledge in bash string input manipulation magic, so any help would be appreciated.

Answer 1

No magic required - save the values in an array when they're seen in the input and print them when you hit a --- line:

$ awk -F'=' -v OFS=',' '/---/{print v["fruit"], v["car"]} {v[$1]=$2}' file
apple,bmw
orange,ford