Mongo 文本搜索多个短语 - AND、OR

Question

以下内容返回所有包含短语“Comp and Comp”和“低威胁”的文档

{ $text: { $search: '"Comp and Comp" "Low Threat"' } }

同理，如何对多个短语进行 OR 文本搜索？

我能得到的最近的信息如下：

{ $text: { $search: 'Comp and Comp Low Threat' } }

但它没有办法对短语进行分组，并且仅对单个字符串执行 OR 操作。

Answer 1

遗憾的是，由于

$text

限制

，您必须执行两个单独的调用

一个查询最多可以指定一个$text表达式。

如果 $search 字符串包含短语和单个术语，则文本搜索将仅匹配包含该短语的文档。

本质上，两个短语在它们之间执行

$and

逻辑，并且因为您只能执行单个

$text

查询，所以需要将其拆分为 2 个调用。

Answer 2

这是该问题的解决方案（对术语和短语列表进行OR搜索）（PyMongo）。

def search_terms_or(collection, terms, projection={}):
    """
    Perform an OR search over a list of terms, which can be unigrams (single words) or phrases (multiple words),
    combining the results while avoiding duplicates.

    Parameters:
    - terms (list of str): A list of search terms, which can be unigrams or phrases (with spaces).
    - projection (dict, optional): MongoDB projection specification to control which fields to return.
      Default is an empty dict, which returns all fields.

    Returns:
    - list: Combined list of unique documents matching any of the search terms.
    """

    # split terms into unigrams and phrases
    unigrams = [term for term in terms if " " not in term]
    phrases = [term for term in terms if " " in term]

    results = []
    seen_ids = set()

    # search for unigrams, if any
    if unigrams:
        unigram_search_string = " ".join(unigrams)
        unigram_results = collection.find({"$text": {"$search": unigram_search_string}}, projection)
        results = list(unigram_results)
        seen_ids = {doc['_id'] for doc in results}
    
    # search for each bigram and add new results
    for phrase in phrases:
        phrase_results = collection.find({"$text": {"$search": f'\"{phrase}\"'}}, projection)
        for doc in phrase_results:
            if doc['_id'] not in seen_ids:
                results.append(doc)
                seen_ids.add(doc['_id'])
    
    return results

然后调用

search_terms_or(myCollection, ["Comp and Comp", "Low Threat"])

会给你你想要的结果。