如何使用jq在字符串中漂亮地打印JSON？

我正在用jq编写一个更复杂的转换，我想做的一件事是在字符串中有一个漂亮的打印JSON。例如：

echo '{"foo": "bar"}' | jq '{json: {other: .} | tostring}'

给予

虽然我想得到：

{
  "json": "{\n  \"other\": {\n    \"foo\": \"bar\"\n  }\n}"
}

我也尝试过

tojson

和

@json

，但它们给出的结果与

tostring

相同。有可能用jq吗？或者我必须求助于其他的诡计吗？请注意，我需要在输出中包含多个带格式JSON的字符串，而不是像示例中那样的字符串。

echo '{"foo": "bar"}' | jq '{other: .}' | jq -Rs '{json: .}'

产生：

{
  "json": "{\n  \"other\": {\n    \"foo\": \"bar\"\n  }\n}\n"
}

删除终止的

“\n”

的一种方法是将其剥离：

echo '{"foo": "bar"}' | jq '{other: .}' | jq -Rs '{json: .[:-1]}'

这：

产生：

{
  "json": "{\n  \"other\": {\n    \"foo\": \"bar\"\n  }\n}\n"
}

删除终止的

“\n”

的一种方法是将其剥离：

echo '{"foo": "bar"}' | jq '{other: .}' | jq -Rs '{json: .[:-1]}'

---

最后我编写了一个简单的格式化函数：

#   9 = \t
#  10 = \n
#  13 = \r
#  32 = (space)
#  34 = "
#  44 = ,
#  58 = :
#  91 = [
#  92 = \
#  93 = ]
# 123 = {
# 125 = }

def pretty:
  explode | reduce .[] as $char (
    {out: [], indent: [], string: false, escape: false};
    if .string == true then
      .out += [$char]
      | if $char == 34 and .escape == false then .string = false else . end
      | if $char == 92 and .escape == false then .escape = true else .escape = false end
    elif $char == 91 or $char == 123 then
      .indent += [32, 32] | .out += [$char, 10] + .indent
    elif $char == 93 or $char == 125 then
      .indent = .indent[2:] | .out += [10] + .indent + [$char]
    elif $char == 34 then
      .out += [$char] | .string = true
    elif $char == 58 then
      .out += [$char, 32]
    elif $char == 44 then
      .out += [$char, 10] + .indent
    elif $char == 9 or $char == 10 or $char == 13 or $char == 32 then
      .
    else
      .out += [$char]
    end
  ) | .out | implode;

它在空对象和数组中添加了不必要的空行，但对于我来说已经足够好了。例如（单独使用）：

函数保存在

pretty.jq

和

test.json

文件中的位置是：

{"type":"FeatureCollection","features":[{"type":"Feature","properties":{"key":"string with \"quotes\" and \\"},"geometry":{"type":"Polygon","coordinates":[[[24.2578125,55.178867663281984],[22.67578125,50.958426723359935],[28.125,50.62507306341435],[30.322265625000004,53.80065082633023],[24.2578125,55.178867663281984]]]}}]}

给出：

{
  "type": "FeatureCollection",
  "features": [
    {
      "type": "Feature",
      "properties": {
        "key": "string with \"quotes\" and \\"
      },
      "geometry": {
        "type": "Polygon",
        "coordinates": [
          [
            [
              24.2578125,
              55.178867663281984
            ],
            [
              22.67578125,
              50.958426723359935
            ],
            [
              28.125,
              50.62507306341435
            ],
            [
              30.322265625000004,
              53.80065082633023
            ],
            [
              24.2578125,
              55.178867663281984
            ]
          ]
        ]
      }
    }
  ]
}

---

最后我编写了一个简单的格式化函数：

#   9 = \t
#  10 = \n
#  13 = \r
#  32 = (space)
#  34 = "
#  44 = ,
#  58 = :
#  91 = [
#  92 = \
#  93 = ]
# 123 = {
# 125 = }

def pretty:
  explode | reduce .[] as $char (
    {out: [], indent: [], string: false, escape: false};
    if .string == true then
      .out += [$char]
      | if $char == 34 and .escape == false then .string = false else . end
      | if $char == 92 and .escape == false then .escape = true else .escape = false end
    elif $char == 91 or $char == 123 then
      .indent += [32, 32] | .out += [$char, 10] + .indent
    elif $char == 93 or $char == 125 then
      .indent = .indent[2:] | .out += [10] + .indent + [$char]
    elif $char == 34 then
      .out += [$char] | .string = true
    elif $char == 58 then
      .out += [$char, 32]
    elif $char == 44 then
      .out += [$char, 10] + .indent
    elif $char == 9 or $char == 10 or $char == 13 or $char == 32 then
      .
    else
      .out += [$char]
    end
  ) | .out | implode;

它在空对象和数组中添加了不必要的空行，但对于我来说已经足够好了。例如（单独使用）：

函数保存在

pretty.jq

和

test.json

文件中的位置是：

{"type":"FeatureCollection","features":[{"type":"Feature","properties":{"key":"string with \"quotes\" and \\"},"geometry":{"type":"Polygon","coordinates":[[[24.2578125,55.178867663281984],[22.67578125,50.958426723359935],[28.125,50.62507306341435],[30.322265625000004,53.80065082633023],[24.2578125,55.178867663281984]]]}}]}

给出：

{
  "type": "FeatureCollection",
  "features": [
    {
      "type": "Feature",
      "properties": {
        "key": "string with \"quotes\" and \\"
      },
      "geometry": {
        "type": "Polygon",
        "coordinates": [
          [
            [
              24.2578125,
              55.178867663281984
            ],
            [
              22.67578125,
              50.958426723359935
            ],
            [
              28.125,
              50.62507306341435
            ],
            [
              30.322265625000004,
              53.80065082633023
            ],
            [
              24.2578125,
              55.178867663281984
            ]
          ]
        ]
      }
    }
  ]
}

---

如果我需要有多个这样的格式化字符串呢？假设输入是

[{“foo”：“one”}，{“foo”：“two”}，…]

，输出是

[{“other”：“…”}，{“other”：“…”}，…”

。我还需要使用输入JSON的一些属性来创建输出。假设输入是

[{“foo”：“one”，“prop”：“prop one”}，{“foo”：“two”，“prop”：“prop two}，…”

，输出是

[{“other”：（格式化的{other:{foo:one}），“prop”：“prop one”}，{“other”：（格式化的{other:{foo:two}），“prop two}，…）

。我认为在jq中编写一个可以漂亮打印JSON字符串的自定义函数可能是最简单的解决方案。如果我需要多个这样的格式化字符串呢？假设输入是

[{“foo”：“one”}，{“foo”：“two”}，

，输出是

[{“other”：“…”}，{“other”：“…}，…]

。我还需要使用输入JSON的一些属性来创建输出。假设输入是

[{“foo”：“one”，“prop”：“prop one”}，{“foo”：“two”，“prop”：“prop two}，…”

，输出是

[{“other”：（格式化的{other:{foo:one}），“prop one”}，{“other”：（格式化的{other:{foo:two}）”，“prop two:“prop two.”}，…]

。我认为在jq中编写一个可以漂亮地打印JSON字符串的自定义函数可能是最简单的解决方案。谢谢，@kszafran。我修改了您的函数，分别打印空数组和空对象的

[]

和

{}

。谢谢，@kszafran。我修改了您的函数，以打印

[]

和

{}

分别用于空数组和对象。