與匹配字符串的一部分從相同線在bash

Question

更換模式我有按照以下格式的行一個JSON文件：

{"id":13, "url":"https://sub.domain.com/path", "dm":"-", "ip":"192.168.0.1"} 
{"id":14, "url":"sub.domain2.com/?param=value", "dm":"-", "ip":"192.168.0.1"} 
{"id":15, "url":"domain.com/path", "dm":"prefilled.com", "ip":"192.168.0.1"} 

，我需要更換「DM」：「 - 」與從同一行各自的領域得到這個輸出：

{"id":13, "url":"https://sub.domain.com/path", "dm":"sub.domain.com", "ip":"192.168.0.1"} 
{"id":14, "url":"sub.domain2.com/?param=value", "dm":"sub.domain2.com", "ip":"192.168.0.1"} 
{"id":15, "url":"domain.com/path", "dm":"prefilled.com", "ip":"192.168.0.1"} 

與任何的bash命令只對具有「DM」的臺詞：「 - 」以優化的方式將文件超過10K的線長

Answer 1

隨着GNU或OSX sed fo通過-E [R ERE支持：

$ sed -E 's#(.*"url":"([^"]+\/\/)?([^"/]+).*"dm":")-"#\1\3"#' file 
{"id":13, "url":"https://sub.domain.com/path", "dm":"sub.domain.com", "ip":"192.168.0.1"} 
{"id":14, "url":"sub.domain2.com/?param=value", "dm":"sub.domain2.com", "ip":"192.168.0.1"} 
{"id":15, "url":"domain.com/path", "dm":"domain.com", "ip":"192.168.0.1"} 

隨着GNU AWK的第三個參數匹配（）：

$ awk 'match($0,/(.*"url":"([^"]+\/\/)?([^"/]+).*"dm":")-(".*)/,a){$0=a[1] a[3] a[4]} 1' file 
{"id":13, "url":"https://sub.domain.com/path", "dm":"sub.domain.com", "ip":"192.168.0.1"} 
{"id":14, "url":"sub.domain2.com/?param=value", "dm":"sub.domain2.com", "ip":"192.168.0.1"} 
{"id":15, "url":"domain.com/path", "dm":"domain.com", "ip":"192.168.0.1"} 

Answer 2

隨着jq-1.5（最新版本ATM），你可以這樣做：

jq 'if .dm == "-" then .dm = (.url|sub("https?://";"")|sub("/.*";"")) else . end' a.json 

說明：

if .dm == "-" ...   # Runs the following only if .dm exists and it's value is "-" 
.dm=(...)     # Assigns to .dm 
.url|sub("^https?://"; "") # Takes .url and replaces http/https:// from the beginning 
...|sub("/.*"; "")   # Replaces everything after the first/(including it) 

Answer 3

您可以用sed來做到這一點，但我建議你使用的東西，實際上領會的數據，如果有格式的任何變化：

sed -i -r 's/^(.*"url":")(.*\/\/)?(.*)(\/.*)"-"/\1\2\3\4"\3"/g' your_file