Memory Used for Input Data
Posted: Wed Apr 11, 2012 5:52 pm
Hello,
I had to change some code element values in a cube. To do this I exported the cube's data, changed the values(update queries), imported the altered data to a new cube.
The issue/mystery is the "Memory Used for Input Data" in the original cube is much larger than it is in the new cube.
Source Cube:
Memory Used for Input Data 1,910,502,400
Number of Populated Numeric Cells 2,550,604
Number of Fed Cells 155,234,934
New Cube:
Number of Populated Numeric Cells 2,552,400
Memory Used for Input Data 225,047,552
Number of Fed Cells 13,344,640
Note: the cube is only 31 MB on disk so I am more comfortable with the new number. The source cube and New cube have the same dimensions. Only the elements of one dimesion were changed.
There is a small variance in the Number of Populated Numeric Cells; this is expected. Since some elements were rollup up to new elements (A, B, C in old cube) is 'D' in the new cube.
The Process:
A) I exported the data from "Cube A" by selecting 'Export to Ascii' with the Skip consolidated Values, Skip Rule Based
Values, and Skip Zero Blank values options selected.
B) Changed a few of the code values by importing the data to a database and running update queries.
C) Imported the file created in step B with a Ti process. No rejects.
I can not account for the large difference in the Memory Used for Input Data.
Thanks,
Walt
I had to change some code element values in a cube. To do this I exported the cube's data, changed the values(update queries), imported the altered data to a new cube.
The issue/mystery is the "Memory Used for Input Data" in the original cube is much larger than it is in the new cube.
Source Cube:
Memory Used for Input Data 1,910,502,400
Number of Populated Numeric Cells 2,550,604
Number of Fed Cells 155,234,934
New Cube:
Number of Populated Numeric Cells 2,552,400
Memory Used for Input Data 225,047,552
Number of Fed Cells 13,344,640
Note: the cube is only 31 MB on disk so I am more comfortable with the new number. The source cube and New cube have the same dimensions. Only the elements of one dimesion were changed.
There is a small variance in the Number of Populated Numeric Cells; this is expected. Since some elements were rollup up to new elements (A, B, C in old cube) is 'D' in the new cube.
The Process:
A) I exported the data from "Cube A" by selecting 'Export to Ascii' with the Skip consolidated Values, Skip Rule Based
Values, and Skip Zero Blank values options selected.
B) Changed a few of the code values by importing the data to a database and running update queries.
C) Imported the file created in step B with a Ti process. No rejects.
I can not account for the large difference in the Memory Used for Input Data.
Thanks,
Walt