基于pandas中的列和字符串计数

Question

假设我有几个文档和一个df列，其中包含我需要搜索的特定单词，如何计算单词出现在文档中的次数？

一个例子更好。

示例：

doc1 = "I am a cat that barks. I like dog food instead of cat food. Roff"

doc2 = "Frog that barks. Frog like cats."

df['words'] = ["dog","cat","frog"]

希望它变成一个看起来像这样的df。

它看起来像这样，但我意识到它只是循环到同一个单元格。所以我总是零。

for i in range(len(doc)):
    for key, value in doc.items():
        for word in df['word']:
            df['doc_' + str(i)] = value.count(word)

Answer 1

doc1 = "I am a cat that barks. I like dog food instead of cat food. Roff"
doc2 = "Frog that barks. Frog like cats."
strings = [doc1, doc2]
words = ["dog","cat","frog"]

def count_occ(word, sentence):
    return sentence.lower().split().count(word)    

cts = []

def counts_df(strings, words):    
    for w in words:
        for s in strings:
            cts.append(count_occ(w, s))
    df = pd.DataFrame(np.array(cts).reshape((len(words), len(strings))),
                      index=words, 
                      columns=['doc' + str(i) for i in range(1, len(strings) + 1)])    
    return df

counts_df(strings, words)
Out[61]: 
      doc1  doc2
dog      1     0
cat      2     0
frog     0     2