Robot Framework data typically uses the whitespace-separated format that looks like this:

*** Test Cases ***
Example
    Keyword    arg1    arg2

In addition to that, we support also the pipe-separated format like:

| *** Test Cases *** |
| Example |
| | Keyword | arg1 | arg2 |

Both of these formats have been supported equally long, since RF 2.1.1 (#375), and they can also be mixed withing a single file because the format is determined based on does a line start with a | or not. In practice the latter syntax is very rarely used and I don't even remember when I have seen it used anywhere except in our tests that validate that the syntax works.

My proposal is to remove the support for the pipe-separated format in order to make the syntax simpler. The main motivation is making it easier to implement new parsers for Robot Framework. It would help making external parsers, for example, for syntax highlighting purposes and it would also make it a lot easier to create a faster parser for Robot itself using C, Rust or Go. The motivation behind #5198 is exactly the same and that issue explains this situation in more detail.

Although this syntax is rarely used, I'm certain it's used somewhere and removing the syntax is obviously backwards incompatible. We thus need to deprecate it first.