Deep parsing

If you have a target data with deep fields - tables or structures - it is possible to fill them in one run. For this you have to create a data source provider implementing zif_text2tab_deep_provider and pass it to the parser. An example of implementation can be found in mockup_loader. But let's consider an simple example here.

Let's assume you have 2 linked tables - header and lines

DOCUMENT
========
ID   DATE   ...
1    ...
2    ...

LINES
========
DOCID   LINEID   AMOUNT   ...
1       1        100.00   ...
1       2        123.00   ...
2       1        990.00   ...

Let's assume you have a target data of the following structure

types:
  begin of ty_line,
    docid  type numc10,
    lineid type numc3,
    " ...
  end of ty_line,
  tt_line type table of ty_line,
  begin of ty_document,
    id   type numc10,
    " ...
    lines type tt_line, " <<< DEEP FIELD, supposed to be filled with lines of the document
  end of ty_document,
  tt_documents type table of ty_document.

So you can run the parser as follows to parse all at once

zcl_text2tab_parser=>create( 
  i_pattern = lt_container 
  i_deep_provider = lo_implementation_of_zif_text2tab_deep_provider
)->parse(
  exporting
    i_data          = my_raw_text_data
  importing
    e_container = lt_container ).

So what is lo_implementation_of_zif_text2tab_deep_provider in this example? The parser does not know how to get additional data other than my_raw_text_data. The implementation of zif_text2tab_deep_provider should. Maybe you want to get it from a file, which is located somewhere near, or download from web, or just select from a DB table - this is up to your design and need. For example, the mentioned mockup_loader uses the text value in place of "deep" field in order to find the data and parse it. The source file should contain a reference in place of "deep" field to help finding the source data. E.g.

DOCUMENT
========
ID   DATE   ...   LINES
1    ...          lines.txt[docid=@id]
2    ...          lines.txt[docid=@id]

where mockup_loader interprets lines.txt[docid=@id] as "go find lines.txt file, parse it, extract the lines, filter those where docid = id value of the current header record".

zif_text2tab_deep_provider requires just one method to implement - select.

methods select
  importing
    i_address type string " e.g. filename[key_field_name=123]
    i_cursor  type any    " reference to currently processed data line to fetch field values
  exporting
    e_container type any.

• i_address is the value of "deep" field in text data (in our example value of LINES field - lines.txt[docid=@id])
• i_cursor is the current record with parsed field values (in our example: id, date, ...)
• e_container is the target component of the current record (in our example - LINES component of ty_document)

So the implementation of the interface must parse the address and get the "deep" data. You are not limited to the specific address format, however, if you want to follow the format described above (<sourceref>[<sourcekeyfield>=<cursorkeyfield>]), you can reuse parse_deep_address and get_struc_field_value_by_name methods implemented in zcl_text2tab_utils.