Sorted table dump in ALV lineformat

[pedrocavaco]

Hi,

Recently I noticed a "strange" dump during some developments.

Here is the detail:

***CODE***

<The code>

lv_day = 1.

WHILE lv_day < lv_datum+6(2).

SHIFT lv_day LEFT DELETING LEADING '0'.

CONCATENATE 'DAY' lv_day INTO lv_fname.

ls_lineformat-fieldname = lv_fname.
ls_lineformat-style = cl_gui_alv_grid=>mc_style_disabled.
APPEND ls_lineformat TO lsx_output-lineformat.
MODIFY gtx_output FROM lsx_output TRANSPORTING lineformat.

lv_day = lv_day + 1.

ENDWHILE.

***DECLARATIONS***

<Just to some context>

ls_lineformat TYPE lvc_s_styl,

lv_fname TYPE lvc_fname,

lv_day TYPE numc2,

lsx_output LIKE LINE OF gtx_output,

gtx_output TYPE STANDARD TABLE OF ty_lpca_mdpd,

TYPES:BEGIN OF ty_lpca_mdpd.
INCLUDE TYPE ty_t_lpca_mdpd.
TYPES: cellcolor      TYPE lvc_t_scol,
lineformat     TYPE lvc_t_styl,
linecolor(4)   TYPE c,
edit TYPE c.
TYPES: END OF ty_lpca_mdpd.

***DUMP***

<The problem>

So far what I noticed was that as long as the creation of DAY1; DAY2; DAY3; ... DAY9; DAY10 stays without a 0 prefix on the number, we get the dump, otherwise if we comment the SHIFT lv_day LEFT DELETING LEADING '0'. and by doing that we get something like DAY01; DAY02; DAY03 ...  DAY09; DAY10 we don't get the dump.

It's strange to me the management which leads the dump, it's supposed?

Thank you for some kind of enlightenment.

[pedrocavaco]

Hi,

Recently I noticed a "strange" dump during some developments.

Here is the detail:

***CODE***

<The code>

lv_day = 1.

WHILE lv_day < lv_datum+6(2).

SHIFT lv_day LEFT DELETING LEADING '0'.

CONCATENATE 'DAY' lv_day INTO lv_fname.

ls_lineformat-fieldname = lv_fname.
ls_lineformat-style = cl_gui_alv_grid=>mc_style_disabled.
APPEND ls_lineformat TO lsx_output-lineformat.
MODIFY gtx_output FROM lsx_output TRANSPORTING lineformat.

lv_day = lv_day + 1.

ENDWHILE.

***DECLARATIONS***

<Just to some context>

ls_lineformat TYPE lvc_s_styl,

lv_fname TYPE lvc_fname,

lv_day TYPE numc2,

lsx_output LIKE LINE OF gtx_output,

gtx_output TYPE STANDARD TABLE OF ty_lpca_mdpd,

TYPES:BEGIN OF ty_lpca_mdpd.
INCLUDE TYPE ty_t_lpca_mdpd.
TYPES: cellcolor      TYPE lvc_t_scol,
lineformat     TYPE lvc_t_styl,
linecolor(4)   TYPE c,
edit TYPE c.
TYPES: END OF ty_lpca_mdpd.

***DUMP***

<The problem>

So far what I noticed was that as long as the creation of DAY1; DAY2; DAY3; ... DAY9; DAY10 stays without a 0 prefix on the number, we get the dump, otherwise if we comment the SHIFT lv_day LEFT DELETING LEADING '0'. and by doing that we get something like DAY01; DAY02; DAY03 ...  DAY09; DAY10 we don't get the dump.

It's strange to me the management which leads the dump, it's supposed?

Thank you for some kind of enlightenment.

[roberto_vacca2]

Hi..

Hands up! It's an illegal statement Well I suppose that if your key is a char key, you have to use 01,02,03 because sort makes a lexicographical order and you need 0 or you'll have something like

1, 10, 11, 12,13,...19, 2, 21...etc...

Hope to help.

[Former Member]

The reason it is giving dump is that append always try to insert the record at the end and it is viaolating you sort order for sorted table lsx_output-lineformat.

Use INSERT ls_lineformat INTO TABLE lsx_output-lineformat.

R

[Former Member]

Rudra has right.

lvc_t_styl is sorted table. Append-statement is invalid for sorted tables. Use insert-statement instead.

[pedrocavaco]

Thank you all for your responses, worked like a charm.