问题描述
我正在尝试找到一种方法来压缩来自sap cats表的数据。我只需要每个“ belnr”一个条目，而不是字段“ belnr”的多余条目。各个猫工时应添加到“ totalhours”的新字段中，并且还应该有一个新的“ counter”字段，该字段显示了总结了多少个相等的“ belnr”条目。 我有不同的解决方案，现在想比较运行时以获得最佳性能。进行一些测试，一切正常。但是，当我添加更多测试条目（大约在6000左右）时，我总是会收到异常“ CX_SY_ARITHMETIC_ERROR”。转换为“ 2995”时会发生。
下面，我向您展示声明部分和一个解决方案，因为所有不同的解决方案都需要大量阅读。希望你们能帮助我。我认为应该很容易解决，但是我无法解决...
TYPES: BEGIN OF ty_cats,counter TYPE catscounte,raufnr TYPE eaufnr,belnr TYPE catsbelnr,kokrs TYPE kokrs,catshours TYPE catshours,END OF ty_cats,tyt_cats TYPE STANDARD TABLE OF ty_cats.
TYPES: BEGIN OF ty_cats_belnr,belnr TYPE catsbelnr,kokrs TYPE kokrs,objnr TYPE j_objnr,count TYPE i,totalhours TYPE catshours,END OF ty_cats_belnr,tyt_cats_belnr TYPE STANDARD TABLE OF ty_cats_belnr.
TYPES: BEGIN OF ty_cats_belnr_sorted,END OF ty_cats_belnr_sorted,tyt_cats_belnr_sorted TYPE SORTED TABLE OF ty_cats_belnr_sorted WITH UNIQUE KEY belnr.
DATA: lt_cats_belnr TYPE tyt_cats_belnr,lt_cats_belnr_sorted TYPE tyt_cats_belnr_sorted,lt_catsdb TYPE tyt_cats,lt_catsdb_buf TYPE tyt_cats,lv_counter TYPE catscounte,lv_totalhours TYPE catshours,lv_start_time TYPE timestampl,lv_end_time TYPE timestampl,lv_diff TYPE p DECIMALS 5.
ParaMETERS: p_enl(3) TYPE n DEFAULT '010'.
START-OF-SELECTION.
*..: internal table with test entries
lt_catsdb = VALUE #( ( counter = 1 kokrs = '1100' belnr = 10001 raufnr = 900000 catshours = '1.20' )
( counter = 2 kokrs = '1100' belnr = 10011 raufnr = 900000 catshours = '2.00' )
( counter = 3 kokrs = '1100' belnr = 10001 raufnr = 900000 catshours = '4.50' )
( counter = 4 kokrs = '1100' belnr = 10012 raufnr = 900000 catshours = '1.00' )
( counter = 5 kokrs = '1100' belnr = 10002 raufnr = 900000 catshours = '1.50' )
( counter = 6 kokrs = '1100' belnr = 10003 raufnr = 900000 catshours = '1.00' )
( counter = 7 kokrs = '1100' belnr = 10002 raufnr = 900000 catshours = '3.50' )
( counter = 8 kokrs = '1100' belnr = 10004 raufnr = 900000 catshours = '1.00' )
( counter = 9 kokrs = '1100' belnr = 10004 raufnr = 900000 catshours = '2.00' )
( counter = 10 kokrs = '1100' belnr = 10004 raufnr = 900000 catshours = '0.50' )
( counter = 11 kokrs = '1100' belnr = 10005 raufnr = 900000 catshours = '1.00' )
( counter = 12 kokrs = '1100' belnr = 10006 raufnr = 900000 catshours = '1.50' ) ).
*..: copy test entries
DO p_enl TIMES.
lt_catsdb_buf = VALUE #( BASE lt_catsdb_buf ( LInes OF lt_catsdb ) ).
ENDDO.
LOOP AT lt_catsdb_buf ASSIGNING FIELD-SYMBOL(<ls_buf>).
lv_counter = lv_counter + 1.
<ls_buf>-counter = lv_counter.
ENDLOOP.
lt_catsdb = VALUE #( ( LInes OF lt_catsdb_buf ) ).
*..: First solution:
TRY.
CLEAR: lt_cats_belnr_sorted[].
LOOP AT lt_catsdb ASSIGNING FIELD-SYMBOL(<ls_catsdb_2>).
INSERT VALUE #( BASE CORRESPONDING ty_cats_belnr_sorted( <ls_catsdb_2> )
objnr = |OR{ <ls_catsdb_2>-raufnr ALPHA = IN }|
count = REDUCE #( INIT i = 0
FOR <count> IN lt_catsdb WHERE ( belnr = <ls_catsdb_2>-belnr )
NEXT i = i + 1 )
totalhours = REDUCE #( INIT total = '0.00'
FOR <cats_belnr> IN lt_catsdb WHERE ( belnr = <ls_catsdb_2>-belnr )
NEXT total = total + <cats_belnr>-catshours )
) INTO TABLE lt_cats_belnr_sorted.
ENDLOOP.
CATCH cx_root INTO lox_root.
CLEAR lv_error_text.
lv_error_text = lox_root->get_text( ).
IF lv_error_text IS NOT INITIAL.
WRITE:/ 'Fehler bei "insert (komprimierte Form)":'.
WRITE:/ lv_error_text.
ENDIF.
ENDTRY.
解决方法
您的问题是结果表的totalhours类型。它的类型为catshours，实际上是一个长度为4的数量字段，因此它不能接受在itab lt_catsdb的大量记录中填充的总小时数。
这是检查此内容的简单测试：
" this var is equal to 12420.00 if p_enl = 600
DATA(sum_of_all_hours) = REDUCE dmbtr( INIT val TYPE dmbtr FOR wa IN lt_catsdb NEXT val = val + wa-catshours ).
" this is equal to 999.99
DATA(max_possible) = cl_abap_exceptional_values=>get_max_value( lv_totalhours ).
ASSIGN max_possible->* TO FIELD-SYMBOL(<max_possible>).
ASSERT sum_of_all_hours < <max_possible>. " <- this assertion will fail
在这里，我们检查可放入totalhours的{{1}}列的最大值，并发现它超过了其数据类型所允许的最大值，因此出现了错误。
解决方案：使用数量更多，长度更长的数量类型，例如BRGEW：
lt_cats_belnr_sorted