CODEWITHSUNDEEP
cxml-parsinglibxml2
foo_l
foo_l

Reputation: 601

blank removal in xml using libxml2

I have the xml file as:

<Config>
    <tlvid id="2">
              <type>10</type>
              <Devid>001b00100000</Devid>
    </tlvid>

    <tlvid id="3">
             <sessionid>abcd123</sessionid>
    </tlvid>

The code which parses the xml file is:

xmlNode *cur_node = NULL,*sub_node = NULL;
xmlChar *key;

cur_node = a_node->xmlChildrenNode;
while(cur_node !=NULL) {

        if((!xmlStrcmp(cur_node->name,(const xmlChar*)"tlvid"))) {
            key = xmlGetProp(cur_node,(const xmlChar*)"id");
            printf("key: %s\n ",key);
            xmlFree(key);
            sub_node = cur_node->xmlChildrenNode;

            while(sub_node !=NULL) {

            key = xmlNodeGetContent(sub_node);
            printf("subkey: %s\n ",key);
            xmlFree(key);

            sub_node = sub_node->next;
            }
        }
     cur_node = cur_node->next;
}

The output as:

key: 2

subkey:

subkey: 10

subkey:

subkey: 001b00100000

subkey:

key: 3

subkey:

subkey: abcd123

subkey:

I have tried xmlKeepBlanksDefault(0); adding under while loop to avoid blanks,but did not help. Can you please help me in removing these empty blanks. Thanks.

Upvotes: 0

Views: 4425

Answers (1)

MvG
MvG

Reputation: 60988

Avoid processing of the text children of cur_node by checking xmlNodeIsText:

for(sub_node = cur_node->xmlChildrenNode;
    sub_node != NULL;
    sub_node = sub_node->next)
{
    if(xmlNodeIsText(sub_node)) continue;
    …
}

As an alternative to skipping all text nodes, you can make sure only to skip blank nodes by using xmlIsBlankNode:

for(sub_node = cur_node->xmlChildrenNode;
    sub_node != NULL;
    sub_node = sub_node->next)
{
    if(xmlIsBlankNode(sub_node)) continue;
    …
}

The two results will differ if there is non-whitespace text directly within a <tlvid> element.

Read the manual for xmlKeepBlanksDefault to find out the conditions required to make the parser ignore these blank nodes. Apparently you need a validating parser and a suitable DTD for this to have any effect.

Upvotes: 2

Related Questions