基于父节点提取 XML 子节点值

Question

我有一个相当大的 xml，看起来像这样：

<?xml version="1.0" encoding="ISO-8859-1"?>
<!DOCTYPE tv SYSTEM "xmltv.dtd">
<tv source-info-url="http://blah blah blah.com/" source-info-name="blah.com" generator-info-name="zap2xml" generator-info-url="zap2xml@gmail.com">
                        <channel id="IX.XXXXXX.blah.com">
                                <display-name>WCBS</display-name>
                                <display-name>2 WCBS</display-name>
                                <display-name>2</display-name>
                                <icon src="https://blah blah blah.png" />
                        </channel>
                        <channel id="IX.XXXXX.blah.com">
                                <display-name>WCBSDT</display-name>
                                <display-name>2 WCBSDT</display-name>
                                <display-name>2</display-name>
                                <icon src="https://blah blah blah.png" />
                        </channel>
                        <channel id="IX.XXXXX.blah.com">
                                <display-name>WNBC</display-name>
                                <display-name>4 WNBC</display-name>
                                <display-name>4</display-name>
                                <icon src="https://blah blah blah.png" />
                        </channel>
                        .....
    </tv>

现在，我只想遍历 .xml，提取频道 ID TAG 值，然后提取第一个和第三个“显示名称”标记值，并根据该 ID 在每个频道显示它们。

我有这个：

#!/bin/bash
file='/path/to/xml/file/file.xml'
cat $file | while read line ; do
        if [[ $line == *"<channel id="* ]]; then
        channelid=$(echo $line|awk -F'"' '{print $2}')
        channelnum=$(xmlstarlet sel -t -v "//channel[@id='$channelid']//display-name[3]" -n $file)
        callsign=$(xmlstarlet sel -t -v "//channel[@id='$channelid']//display-name[1]" -n $file)
        clear
        echo "Here are the details for channel id: $channelid"
        echo ""
        echo "Channel Number is: $channelnum"
        echo "Channel Call Sign is: $callsign"
        echo ""
        sleep 2
        fi
done

which DOES 做我想做的事，但整个过程变慢了，因为它不断尝试查找它遇到的每个通道的外部实体链接 - 输出：

Here are the details for channel id: IX.XXXXX.blah.com

Channel Number is: 2
Channel Call Sign is: WCBS

/path/to/the/epg.xml:2.30: failed to load external entity "/path/to/the/script/xmltv.dtd"
<!DOCTYPE tv SYSTEM "xmltv.dtd">
                                ^
/path/to/the/xmlfile.xml:2.30: failed to load external entity "/path/to/the/script/xmltv.dtd"
<!DOCTYPE tv SYSTEM "xmltv.dtd">

如何抑制这些查找？我只想解析这些值。

Answer 1

如何抑制这些查找？

您可以先使用xmlstarlet fo (format) command 删除 DOCTYPE 声明。

您还应该能够通过使用xmlstarlet sel (select) command 和XPath 来获得您需要的所有值。尝试使用正则表达式对文件进行分类、读取行和解析可能很脆弱。

完整示例...

#!/bin/bash

file='/path/to/the/file.xml'

xmlstarlet fo -D $file | \
xmlstarlet sel -T -t -m "/tv/channel" \
    -v "concat('Here are the details for channel id: ', @id)" -nl -nl \
    -v "concat('  Channel Number is: ', display-name[3])" -nl \
    -v "concat('  Channel Call Sign is: ', display-name[1])" -nl -nl

输出...

Here are the details for channel id: IX.XXXXXX.blah.com

  Channel Number is: 2
  Channel Call Sign is: WCBS

Here are the details for channel id: IX.XXXXX.blah.com

  Channel Number is: 2
  Channel Call Sign is: WCBSDT

Here are the details for channel id: IX.XXXXX.blah.com

  Channel Number is: 4
  Channel Call Sign is: WNBC