How to specify the Flat File Schema to handle random sequence of different line types?

I have a two types of flat file lines,

1. start with @, e.g. @111 222 333

2. start with #, e.g.   #777 888 999

We can use tag identifier to identify these two types of lines. Unfortunately, these two lines may appear in any sequence in a file, like @@###, or @#@#@, or ###@@ etc.

If use flat file dissembler schema, seems the sequence must be in order, say @@@##, @##, it can disassemble all flat file content into xml format.

But if the file format is @@#@@# or @###@#, it cannot dissemble all file content and will stop at the middle, say @@#@@# can only disassemble @@#, @###@# can only disassemble @###.

I think it is because the flat file schema specify the sequence is @ and then #, therefore, @ cannot appear after # (please inform if my understanding is wrong).

Any solution to our case?

Thanks.