提取XML数据的速度比嵌套循环更快？

Question

我有一个具有以下格式的XML文件：

<JMdict>
...
<entry>
        <ent_seq>2232410</ent_seq>
        <k_ele>
                <keb>筆おろし</keb>
        </k_ele>
        <k_ele>
                <keb>筆下ろし</keb>
        </k_ele>
        <k_ele>
                <keb>筆降ろし</keb>
                <ke_inf>&iK;</ke_inf>
        </k_ele>
        <r_ele>
                <reb>ふでおろし</reb>
        </r_ele>
        <sense>
                <pos>&n;</pos>
                <pos>&vs;</pos>
                <gloss>using a new brush for the first time</gloss>
        </sense>
        <sense>
                <pos>&n;</pos>
                <pos>&vs;</pos>
                <gloss>doing something for the first time</gloss>
        </sense>
        <sense>
                <pos>&n;</pos>
                <pos>&vs;</pos>
                <gloss>man losing his virginity (esp. to an older woman)</gloss>
        </sense>
</entry>
...
</JMdict>

链接到整个XML文件：

这基本上是电子日语/英语词典。有很多入口标签。我正在尝试创建一个搜索函数，该功能将根据任何keb，reb和“光泽”标签中的文本值返回ent_seq号码。

我有鲍泽代码可以执行我需要做的事情，但有些慢（438毫秒）。然后，该SEQ号码将用于在另一个数据集中查找数据，如果我打算在Web应用中使用它，我希望它更快。有办法吗？

from xml.etree import ElementTree as ET

tree = ET.parse("../../resources/JMdict_e.xml")
root = tree.getroot()

search_term = '筆おろし'
seq_tags = []

for dictionary in root.iter('JMdict'):
    
    for child in dictionary:
        
        for grandchild in child:
            if grandchild.tag == 'ent_seq':
                ent_seq = grandchild.text
                
            for greatgrandchild in grandchild:
                if greatgrandchild.tag in ['keb','reb','gloss']:
                    if greatgrandchild.text == search_term:
                        seq_tags.append(ent_seq)

                    
print(seq_tags)

任何帮助和技巧将不胜感激。

Answer 1

使用XPATH在同一表达式中搜索超过1个元素（可能是2个以上条件的一个和/或表达式）

from lxml import etree
import time
import io

st = time.process_time()
parser = etree.XMLParser(compact=True, huge_tree=True, resolve_entities=False)
with open('/home/luis/tmp/JMdict_e', 'rb') as f:
    et = time.process_time()
    res = et - st
    print('read file:', res, 'seconds')
    tree = etree.parse(f, parser)
    
    et = time.process_time()
    res1 = et - res
    print('parse:', res1, 'seconds')

    slist = tree.xpath('//entry[k_ele/keb = "筆おろし"]/ent_seq | //entry[r_ele/reb = "エヌきょう"]/ent_seq')
    #slist = tree.xpath('//entry[1]/k_ele/keb')
    et = time.process_time()
    res2 = et - res1
    print('xpath:', res, 'seconds')
    
    #print(slist)
    for d in slist:
        print( d.text)

et = time.process_time()
res = et - st
print('CPU Execution time:', res, 'seconds')