Identify the TYPE using CL_ABAP_DATADESCR

Question

we have a custom field ROLE  with the Data element as ZDOA_ROLE in a table ZTABLE.
CODE SAMPLE 1
data gs_comp          TYPE abap_componentdescr.
gs_comp-type ?=  cl_abap_datadescr=>describe_by_data( 'ROLE' ).
gs_comp-name = gs_comp-type->get_relative_name( ).
when i write the above code,  the name is blank.
CODE SAMPLE 2
do_type =  cl_abap_datadescr=>describe_by_name( 'ROLE' ).
it dumps, stating ROLE type is not available
CODE SAMPLE 3
do_type =  cl_abap_datadescr=>describe_by_name( 'ZDOA_ROLE' ).
it returns the type is ZDOA_ROLE.            
"--NOT USEFUL because i cannot pass the Data element, I can only pass Field name
Priority is to find the type from the Field.
Can any body suggest me, a better way to get the Type of the Field and not by the Data element ?

Answer 1

abhishek nms wrote:
Any work around ? because i cannot mention the table name,
abhishek nms wrote:
Requirement: based on the tab delimited file upload in Web Dynpro, I need to identify which table data (mostly our Custom ZDOA tables which are around 10) was uploaded and updated the corresponding table .
These two statements seem contradictory. If you know that the file must be one of your ZDOA tables, than you can get the table name(s) by finding all tables that match the pattern ZDOA*
Once you know the possible tables, you can determine the structure of each table like this:
lo_structdescr ?= cl_abap_structdescr=>describe_by_name( 'ZDOA_TABLE' ).
li_components = lo_structdescr->get_components( ).
li_components is a table of the names and type descriptors for each field in the structure. You can loop through this table, comparing the names of each field to the field names given in the header of the tab-delimited file.