请教下如何提取 curl 格式化输出的值

V2EX = way to explore

V2EX 是一个关于分享和探索的地方

现在注册

已注册用户请登录

Distributions

› Ubuntu

› Fedora

› CentOS

中文资源站

› 网易开源镜像站

这是一个创建于 2649 天前的主题，其中的信息可能已经有所发展或是发生改变。

如题：想把 curl 得到的值按照自己想要的格式输出到文本，方便监控
curl 输出的值：
{"success":true,"data":{"request_5xx":1,"traffic_write":10866877,"average_traffix_write":426.43633010242,"average_traffic_read":429.33143664404,"traffic_read":10940653,"average_request_time":0.44471063102306,"total_count":25483,"request_3xx":0,"request_2xx":23981,"request_4xx":1501,"total_request_time":11332.561010361}
想得到的格式：
traffic_write:10866877
request_4xx:1501

curl

输出

格式

success

15 条回复 • 2017-08-23 14:26:01 +08:00

Chyroc

2017-08-22 10:41:39 +08:00

echo '{"success":true,"data":{"request_5xx":1,"traffic_write":10866877,"average_traffix_write":426.43633010242,"average_traffic_read":429.33143664404,"traffic_read":10940653,"average_request_time":0.44471063102306,"total_count":25483,"request_3xx":0,"request_2xx":23981,"request_4xx":1501,"total_request_time":11332.561010361}}' | jq '. | {traffic_write: .data.traffic_write, request_4xx: .data.request_4xx}'

Chyroc

2017-08-22 10:42:06 +08:00

把 echo '{"success":true,"data":{"request_5xx":1,"traffic_write":10866877,"average_traffix_write":426.43633010242,"average_traffic_read":429.33143664404,"traffic_read":10940653,"average_request_time":0.44471063102306,"total_count":25483,"request_3xx":0,"request_2xx":23981,"request_4xx":1501,"total_request_time":11332.561010361}}' 换成 curl xxxx 即可

zhs227

2017-08-22 10:49:02 +08:00

yum install jq
https://stedolan.github.io/jq/manual/

jq 比较好用。下面是一个例子：
https://jqplay.org/s/vsOpssDUO_

krisbai

2017-08-22 10:58:45 +08:00

@zhs227 嗯，这个工具确实蛮好用的，之前也在用，想请教下我格式化出来的带逗号，这个如何在获取参数的时候取消掉？
比如：cat page.txt |jq .|grep request_4xx| cut -d ":" -f 2 得到的值：1505，想取消掉逗号

zhs227

2017-08-22 11:00:27 +08:00

试试 `jq --raw-output`

zhs227

2017-08-22 11:02:11 +08:00

用 jq 直接取字段，不要用 grep 处理。1 楼的答案。比如 jq ".data.request_4xx"

NoAnyLove

2017-08-22 11:12:33 +08:00

话说直接用命令行下的 tr 和 grep 就行了啊，不需要安装额外的工具：

echo '{"request_5xx":1,"traffic_write":10866877,"average_traffix_write":426.43633010242,"average_traffic_read":429.33143664404,"traffic_read":10940653,"average_request_time":0.44471063102306,"total_count":25483,"request_3xx":0,"request_2xx":23981,"request_4xx":1501,"total_request_time":11332.561010361}' | tr -d '"' | grep -oP '(traffic_write:\d+)|(request_4xx:\d+)'

krisbai

2017-08-22 11:20:04 +08:00

@Chyroc 感谢

krisbai

2017-08-22 11:20:26 +08:00

@zhs227 哈哈，不知道我格式有问题还是，得到的数据有点问题

krisbai

2017-08-22 11:20:40 +08:00

@NoAnyLove 谢谢大牛，这个办法真直接！

krisbai

2017-08-22 13:16:32 +08:00

@NoAnyLove 再请教下，生成的数据都在一行。如果要达到换行的效果，\n 应该加在那里呢？

ms2008

2017-08-22 15:29:26 +08:00

```
echo '{"success":true,"data":{"request_5xx":1,"traffic_write":10866877,"average_traffix_write":426.43633010242,"average_traffic_read":429.33143664404,"traffic_read":10940653,"average_request_time":0.44471063102306,"total_count":25483,"request_3xx":0,"request_2xx":23981,"request_4xx":1501,"total_request_time":11332.561010361}}' | python -mjson.tool
```

UnknownR

2017-08-22 20:32:04 +08:00 via iPhone

输出成 json 或者 xml 就可以了，都是数据存储格式，会自动调整的

NoAnyLove

2017-08-23 13:18:40 +08:00

@krisbai 显示成一行了？我这里显示的是两行啊，你是要把两行变成一行？还是一行变成两行？反正这两种都可以通过 sed 实现。

两行内容变成一行：
echo 'traffic_write:10866877\nrequest_4xx:1501' | sed 'N;s/\n/\t/g'

一行内容变成两行：
echo 'traffic_write:10866877\trequest_4xx:1501' | sed 's/\s\+/\n/g'

把对应的 sed 命令继续用管道接在之前的命令后面就行了。

krisbai

2017-08-23 14:26:01 +08:00

@NoAnyLove 好，谢谢！