与jq匹配的子树

Question

我正在寻找一种有效的方法来计算一个 JSON 文档是否是另一个 JSON 文档的子树/子集。 给定一个 JSON 文档 a

{
    "a": {
        "a": {"aa": ["uvw", "abc"]},
        "b": {
            "ba": {
                "baa": ["none", "1"],
                "bab": ["0"]
            },
            "bb": {
                "baa": ["yyy"],
                "bab": ["some_string"]
            }
        },
        "c": {
            "ca": ["dd", "cc"],
            "cb": 2,
            "cc": "-" 
        }
    },
    "e": "abc",
    "d": 45678
}

还有一个 JSON 文档 b

{
    "a": {
        "a": {"aa": ["abc"]},
        "b": {
            "ba": {
                "baa": ["none"],
                "bab": ["0"]
            },
            "bb": {
                "bab": ["some_string"]
            }
        },
        "c": {
            "ca": ["cc"]
        }
    },
    "e": "abc"
}

我想确定b 是否是a 的子集

因此，我希望有一个布尔值（或类似 {'match': true} 的东西）表明 b 是 a 的子集。

我尝试了以下

.[0] as $a | .[1] as $b
| reduce ($a|paths) as $p (null; ($a|getpath($p)) as $va
| try($b|getpath($p)) as $vb
| if $va == $vb then setpath(["match"];true) else setpath(["match"];false) end)

但我错过了跟踪失败比较的方法。

如何使用jq 完成这种子树匹配？

提前感谢您的帮助！

Answer 1

正如@weeble 所指出的，当涉及到数组时，“子树”的概念并不是那么简单。特别是，针对您的问题的简单路径值方法不符合您的要求，因此以下内容有点复杂，但希望布局和命名法有助于澄清问题。

首先，一些辅助函数：

def max(s): reduce s as $x (0; if . == null or $x > . then $x else . end);

def elementOf($x):
  . as $in
  | if ($x|type) == "array"
    then $x|index($in)
    else false end;

# drill down
def shift($n):
  if $n == 0 then .
  elif type | (. != "object" and . != "array") then empty
  elif $n == 1 then .[]
  else .[] | shift($n - 1)
  end;

接下来是主函数，它定义了一个（递归的）子函数来处理 相对简单的案例：

def subtree_of($b):
  def isScalar: type | . != "array" and . != "object";

  # direct_subtree_of($b) checks that every path/atomic-value pair of $a is also in $b
  def direct_subtree_of($b):
    . as $a
    | first( ($a|paths(scalars)) as $p
        | ($a | getpath($p)) as $va
        | ([try ($b|getpath($p)) // empty]) as $vb
        | if ($vb |length > 0) 
             and (($va == $vb[0])
                  or ($va | subtree_of($vb[0]))
                  or ($va | elementOf($b | getpath( $p[:-1] ))))
          then empty
          else 0
          end) // 1
      | . == 1 ;
    
  . as $a
  | if isScalar then $a == $b
    else direct_subtree_of($b)
         or (max($a|paths|length) as $ma
            | max($b|paths|length) as $mb
            | ($mb - $ma) as $diff
            | any(range(1; 1+$diff);
                  any(. as $i | $a | subtree_of($b | shift($i)); . )) )
    end ;

更正你的a后：

a | subtree_of(b)

计算为true。