Hey its me again
<br />
<br />
I think this one is not possible, but if it would be great.<br />
If I have one DataSet where there are Orders for all products in a specific timespan. To every order I have calculated a Yield. Now I want to do the following:<br />
Give to each product the last four (not more and not less) orders with orderNo and Yield each (lets call them YIELD_1, YIELD_2, YIELD_3, YIELD_4). Out of these Yields calculate a trend of two following ones. Something like:<br />
<pre class='_prettyXprint _lang-auto _linenums:0'>
Yieldij =YIELD_j - YIELD_i
</pre>
Then do a visibility filter on the list where you say:<br />
<pre class='_prettyXprint _lang-auto _linenums:0'>
if((row["Yield12"]<0 && row["Yield23"]<0)||(row["Yield23"]<0 && row["Yield34"]<0)){
false;}
else {
true;}
</pre>
<br />
I have no clue how to do it and I think it isn't possible. My solution so far is to create two DataSets. One where are all the ProductNOs and -descriptions and a second that gives you for one Product the last four orders in one ROW(big JOIN
).<br />
The I created a table where BIRT should give me the ProductNOs and -descriptions and created another table within the detail field of the first table with the PrNO of each row as parameter that gives me the last four orders. Then I calculate Yield12, Yield23, Yield34 and applied the visibility above.<br />
<br />
The problems are:<br />
The delay is much to long!!!<br />
I can't apply the visibility filter to the whole table but only the sub-tables, so the report gives me ALL products and the subtable only for the interesting ones.<br />
<br />
Maybe someone has an idea how to do it. I thought of JavaScript temp variables, but because everything is in one data Set I don't think that will work...
<br />
<br />
Thank you very much!<br />
Best regards,<br />
JRS
