Untitled

def func_sqlcol( __pv_df_param, __pdf_large_cols):
    __dtypedict = {}
    __cols_clob = []
    __cols_blob = []
    try:
        #print(("__df_large_cols: \n%s" % __pdf_large_cols))
        if len(__pdf_large_cols) > 0:
            __cols_clob = __pdf_large_cols['column_name_clob'].values[0]
            __cols_clob = __cols_clob.split(',') if __cols_clob else []
            
            __cols_blob = __pdf_large_cols['column_name_blob'].values[0]
            __cols_blob = __cols_blob.split(',') if __cols_blob else []
            
        print(("__cols_clob : type{%s} len{%s}" % (type(__cols_clob), len(__cols_clob))))
        print(("__cols_blob : type{%s} len{%s}" % (type(__cols_blob), len(__cols_blob))))
        
        for i,j in zip(__pv_df_param.columns, __pv_df_param.dtypes):
            # table(s): additionaldet_xml_template.xml_template, tom_batch_order_template_ref.template
            if "object" in str(j) and i in __cols_blob:
                __dtypedict.update({i: sqlalchemy_types.BLOB(length=None)})
            elif "object" in str(j) and i in __cols_clob:
                # table(s): tom_service_action_master.extended_rule, generic_template_ref.template_text
                __dtypedict.update({i: sqlalchemy_types.CLOB(length=None)})
            elif "object" in str(j): 
                __dtypedict.update({i: sqlalchemy_types.NVARCHAR(length=255)})
            
            if "datetime" in str(j):
                __dtypedict.update({i: sqlalchemy_types.DateTime()})
            
            if "float" in str(j):
                __dtypedict.update({i: sqlalchemy_types.Float(precision=3, asdecimal=True)})
            
            if "int" in str(j):
                __dtypedict.update({i: sqlalchemy_types.INT()})
    except Exception as __exp:
        print(("func_sqlcol exception. Error - {%s} . Line No - {%s} " % (str(__exp),str(sys_exc_info()[-1].tb_lineno))))
        raise AirflowException(str(__exp))
   
    return __dtypedict


why do we write above function