Go 每日一庫之 jj

時間  2020-03-27
標籤 每日 简体版
原文   原文鏈接

簡介

在前面兩篇文章中，咱們分別介紹了快速讀取 JSON 值的庫gjson和快速設置 JSON 值的庫sjson。今天咱們介紹它們的做者tidwall的一個基於gjson和sjson的很是實用的命令行工具jj。它是使用 Go 編寫的快速讀取和設置 JSON 值的命令行程序。git

快速使用

Mac 上能夠直接使用brew install tidwall/jj/jj安裝。其餘系統能夠經過下載編譯好的可執行程序，下載地址爲https://github.com/tidwall/jj/releases。github

我選擇使用go get安裝：golang

$ go get github.com/tidwall/jj/cmd/jj

上面命令執行完成以後，編譯生成的jj程序會放在$GOPATH/bin目錄中，我習慣把$GOPATH/bin加入系統可執行目錄$PATH中，故能夠直接使用。shell

簡單的讀取和設置（個人環境爲 Win10 + Git Bash）：json

$ echo '{"name":{"first":"li","last":"dj"}}' | jj name.last
dj

$ echo '{"name":{"first":"li","last":"dj"}}' | jj -v dajun name.last
{"name":{"first":"li","last":"dajun"}}

經過鍵路徑來指定讀取/設置的位置，上面第一個命令讀取字段name.last，返回dj。數組

-v選項指定設置的值。第二個命令將字段name.last設置爲dajun，輸出設置以後的 JSON 串。鍵路徑在前兩篇文章中有詳細的介紹，不熟悉的能夠回去看一下。微信

讀取和設置

實際上讀取和設置的語法和形式與咱們前面介紹gjson和sjson提到的基本同樣，只不過是在命令行上完成的而已。app

讀取不存在的字段，返回null：工具

$ echo '{"name":{"first":"li","last":"dj"}}' | jj name.middle
null

讀取一個對象類型的字段，返回該對象的 JSON 表示：性能

$ echo '{"name":{"first":"li","last":"dj"}}' | jj name
{"first":"li","last":"dj"}

使用索引（從 0 開始）讀取數組的元素，非法的索引將返回空：

$ echo '{"fruits":["apple","orange","banana"]}' | jj fruits.1
orange

$ echo '{"fruits":["apple","orange","banana"]}' | jj fruits.3

使用索引設置數組的元素，下面命令將數組fruits的第二個元素設置爲pear：

$ echo '{"fruits":["apple","orange","banana"]}' | jj -v pear fruits.1
{"fruits":["apple","pear","banana"]}

使用-1或數組長度做爲索引，能夠在數組後添加一個元素。若是索引超過了數組長度，則會多必定數量的null：

$ echo '{"fruits":["apple","orange","banana"]}' | jj -v strawberry fruits.-1
{"fruits":["apple","orange","banana","strawberry"]}

$ echo '{"fruits":["apple","orange","banana"]}' | jj -v grape fruits.3
{"fruits":["apple","orange","banana","grape"]}

$ echo '{"fruits":["apple","orange","banana"]}' | jj -v watermelon fruits.5
{"fruits":["apple","orange","banana",null,null,"watermelon"]}

使用選項-D刪除指定鍵路徑上的元素，若是對應元素不存在，則無效果：

$ echo '{"name":"dj","age":18}' | jj -D age
{"name":"dj"}

$ echo '{"fruits":["apple","orange","banana"]}' | jj -D fruits.2
{"fruits":["apple","orange"]}

$ echo '{"fruits":["apple","orange","banana"]}' | jj -D fruits.5
{"fruits":["apple","orange","banana"]}

第 1 個命令刪除字段age；第 2 個命令刪除數組fruits的第 2 個元素；第 3 個命令刪除數組fruits的第 5 個元素，因爲數組長度只有 3，故無效果。

文件

jj支持從文件中讀取 JSON 串和將結果寫到文件中。使用選項-i指定輸入文件，選項-o指定輸出文件。下面將從文件fruits.txt中讀取 JSON 串，取數組的第 2 個元素，寫到out.txt中：

$ jj -i fruits.txt -o out.txt fruits.1

fruits.txt的文件內容以下：

{"fruits":["apple","orange","banana"]}

執行命令，輸出文件的內容爲：

orange

格式化

jj支持將輸出的 JSON 串進行必定的格式化。選項-u移除全部的空白符，節省存儲空間。選項-p美化格式，便於閱讀。

$ echo '{"name":{"first": "li", "last":"dj"}, "age":18}' | jj -u name
{"first":"li","last":"dj"}

$ echo '{"name":{"first": "li", "last":"dj"}, "age":18}' | jj -p name
{
  "first": "li",
  "last": "dj"
}

性能

與另外一個 JSON 的命令行工具jq相比，jj是其性能的 10 倍以上。由於jj不會驗證 JSON 串的有效性，而且它只關心鍵路徑指定的值，一旦該值處理完成就中止。這裏有性能對比：https://github.com/tidwall/jj#performance

用途

jj一個很方便的用途在於日誌處理，當前不少日誌庫都支持 JSON 的格式，例如前面咱們介紹的logrus。咱們可使用jj在這些日誌中找到相應的信息。咱們先用logrus生成 20 條玩家登錄和下線的日誌：

package main

import "github.com/sirupsen/logrus"

func main() {
  logrus.SetFormatter(&logrus.JSONFormatter{})

  for i := 1; i <= 10; i++ {
    logrus.WithFields(logrus.Fields{
      "userid": i,
    }).Info("login")
    logrus.WithFields(logrus.Fields{
      "userid": i,
    }).Info("logoff")
  }
}

生成日誌存儲在log.txt文件中：

{"level":"info","msg":"login","time":"2020-03-26T23:36:04+08:00","userid":1}
{"level":"info","msg":"logoff","time":"2020-03-26T23:36:04+08:00","userid":1}
{"level":"info","msg":"login","time":"2020-03-26T23:36:04+08:00","userid":2}
{"level":"info","msg":"logoff","time":"2020-03-26T23:36:04+08:00","userid":2}
{"level":"info","msg":"login","time":"2020-03-26T23:36:04+08:00","userid":3}
{"level":"info","msg":"logoff","time":"2020-03-26T23:36:04+08:00","userid":3}
{"level":"info","msg":"login","time":"2020-03-26T23:36:04+08:00","userid":4}
{"level":"info","msg":"logoff","time":"2020-03-26T23:36:04+08:00","userid":4}
{"level":"info","msg":"login","time":"2020-03-26T23:36:04+08:00","userid":5}
{"level":"info","msg":"logoff","time":"2020-03-26T23:36:04+08:00","userid":5}
{"level":"info","msg":"login","time":"2020-03-26T23:36:04+08:00","userid":6}
{"level":"info","msg":"logoff","time":"2020-03-26T23:36:04+08:00","userid":6}
{"level":"info","msg":"login","time":"2020-03-26T23:36:04+08:00","userid":7}
{"level":"info","msg":"logoff","time":"2020-03-26T23:36:04+08:00","userid":7}
{"level":"info","msg":"login","time":"2020-03-26T23:36:04+08:00","userid":8}
{"level":"info","msg":"logoff","time":"2020-03-26T23:36:04+08:00","userid":8}
{"level":"info","msg":"login","time":"2020-03-26T23:36:04+08:00","userid":9}
{"level":"info","msg":"logoff","time":"2020-03-26T23:36:04+08:00","userid":9}
{"level":"info","msg":"login","time":"2020-03-26T23:36:04+08:00","userid":10}
{"level":"info","msg":"logoff","time":"2020-03-26T23:36:04+08:00","userid":10}

因爲每一行都是一個單獨的 JSON 串，咱們可使用jj支持的 JSON 行特性，使用..路徑標識這些行。..使得jj將這些行當作數組的元素。咱們能夠讀取這些數組元素。

獲取數組長度，返回 20：

$ jj -i log.txt ..#
20

只讀取每一行中的userid信息：

$ jj -i log.txt ..#.userid
[1,1,2,2,3,3,4,4,5,5,6,6,7,7,8,8,9,9,10,10]

只讀取每一行中的msg信息：

$ jj -i log.txt ..#.msg
["login","logoff","login","logoff","login","logoff","login","logoff","login","logoff","login","logoff","login","logoff","login","logoff","login","logoff","login","logoff"]

更復雜一點的，若是咱們要查看全部userid=1的日誌：

$ jj -i log.txt ..#\(userid=1\)# -p
[  
  {
    "level": "info",
    "msg": "login",
    "time": "2020-03-26T23:36:04+08:00",
    "userid": 1
  },
  {
    "level": "info",
    "msg": "logoff",
    "time": "2020-03-26T23:36:04+08:00",
    "userid": 1
  }
]

上面的命令注意兩點，(和)是 shell 中的特殊字符，須要\轉義。命令中咱們使用-p選項使結果更易讀。

若是咱們只須要查找第一條符合條件的日誌，則能夠去掉最右側的#：

$ jj -i log.txt ..#\(userid=1\) -p
{
  "level": "info",
  "msg": "login",
  "time": "2020-03-26T23:36:04+08:00",
  "userid": 1
}

若是要查看全部的登陸信息：

$ jj -i log.txt ..#\(msg="login"\)# -p
[
  {
    "level": "info",
    "msg": "login",
    "time": "2020-03-26T23:36:04+08:00",
    "userid": 1
  },
  {
    "level": "info",
    "msg": "login",
    "time": "2020-03-26T23:36:04+08:00",
    "userid": 2
  },
  {
    "level": "info",
    "msg": "login",
    "time": "2020-03-26T23:36:04+08:00",
    "userid": 3
  },
  {
    "level": "info",
    "msg": "login",
    "time": "2020-03-26T23:36:04+08:00",
    "userid": 4
  },
  {
    "level": "info",
    "msg": "login",
    "time": "2020-03-26T23:36:04+08:00",
    "userid": 5
  },
  {
    "level": "info",
    "msg": "login",
    "time": "2020-03-26T23:36:04+08:00",
    "userid": 6
  },
  {
    "level": "info",
    "msg": "login",
    "time": "2020-03-26T23:36:04+08:00",
    "userid": 7
  },
  {
    "level": "info",
    "msg": "login",
    "time": "2020-03-26T23:36:04+08:00",
    "userid": 8
  },
  {
    "level": "info",
    "msg": "login",
    "time": "2020-03-26T23:36:04+08:00",
    "userid": 9
  },
  {
    "level": "info",
    "msg": "login",
    "time": "2020-03-26T23:36:04+08:00",
    "userid": 10
  }
]

總結

jj是一個很是使用的 JSON 命令行工具，性能超讚。執行jj -h去看看其餘選項吧。

你們若是發現好玩、好用的 Go 語言庫，歡迎到 Go 每日一庫 GitHub 上提交 issue😄

參考

  1. jj GitHub：https://github.com/tidwall/jj
  2. Go 每日一庫 GitHub：https://github.com/darjun/go-daily-lib

我

個人博客：https://darjun.github.io

歡迎關注個人微信公衆號【GoUpUp】，共同窗習，一塊兒進步~

相關文章
相關標籤/搜索
每日一句
    每一个你不满意的现在，都有一个你没有努力的曾经。
本站公眾號
   歡迎關注本站公眾號,獲取更多信息
聯繫我們 最近搜索 最新文章 沪ICP备13005482号-10 MyBatis教程 SQL 教程 MySQL教程 Java 教程 Thymeleaf 教程 Hibernate教程 Spring教程 Redis教程