Support indentation of generated xml / html #446

Activity

1. jennybc commented on Sep 6, 2024

   

   jennybc

   on Sep 6, 2024

   Member

   More actions

   By analogy with JSON, I'm interested in taking something like this:

   {"user":{"id":1,"name":"John Doe","contact":{"email":"[email protected]","phone":{"home":"555-555-5555","work":"555-555-5556"}},"address":{"street":"123 Main St","city":"Anytown","state":"CA","postalCode":"12345","coordinates":{"latitude":37.7749,"longitude":-122.4194}},"preferences":{"notifications":{"email":true,"sms":false},"theme":"dark"}}}

   and prettifying it like this:

   json |> jsonlite::prettify()
   #> {
   #>     "user": {
   #>         "id": 1,
   #>         "name": "John Doe",
   #>         "contact": {
   #>             "email": "[email protected]",
   #>             "phone": {
   #>                 "home": "555-555-5555",
   #>                 "work": "555-555-5556"
   #>             }
   #>         },
   #>         "address": {
   #>             "street": "123 Main St",
   #>             "city": "Anytown",
   #>             "state": "CA",
   #>             "postalCode": "12345",
   #>             "coordinates": {
   #>                 "latitude": 37.7749,
   #>                 "longitude": -122.4194
   #>             }
   #>         },
   #>         "preferences": {
   #>             "notifications": {
   #>                 "email": true,
   #>                 "sms": false
   #>             },
   #>             "theme": "dark"
   #>         }
   #>     }
   #> }
   #> 

   But with HTML.
2. hadley

   added

   featurea feature request or enhancementa feature request or enhancement

   on Oct 21, 2024
3. jeroen commented on Jul 3, 2025

   

   jeroen

   on Jul 3, 2025

   Last edited by jeroen

   Member

   More actions

   Seemingly, we already do this by default by enabling NOBLANKS in the reader and format in the writer options. And indeed sometimes it works:

   read_xml('<html><head><title>Yolo</title></head></html>') |> write_xml('output.xml')
   readBin('output.xml', raw(), 1e6) |> rawToChar() |> cat()

   However more often it only adds linebreaks, but not indentation but maybe this is to be expected. I think part of the issue is that whitespace in XML is not ignorable. The libxml2 FAQ says:

   Libxml2 will not invent spaces in the content of a document since all spaces in the content of a document are significant.

   For example, if an xml node contains some code syntax, adding spaces in front can change the code (python for sure 🙄 ) .

   But also note that for example html code <span>Foo</span><b>bar<b> changes meaning when we add whitespace between the nodes because this will become an actual space in the html output try it.