Analytics Catalog/Oracle EPM/FCCS/Retrieval Performance
Explore the catalogFCCSSmart View performanceView dimensionConsolidated Data Export
Oracle EPM Cloud · FCCS · Guidance

Retrieval performance, the FCCS causes

When an FCCS grid crawls, it is almost always asking the cube to compute instead of read. Four causes account for most tickets, and each has a fix that works the same day.

The general diagnosis method, the Activity Report and the worst-query list, lives on Smart View performance. This page carries only what is FCCS's own, and the whole reference is mapped on the FCCS index.

RulePoint grids at stored, level-zero members and let subtotals be subtotals in the sheet. Reads are fast; retrieve-time computation is what you are waiting for.
Neverleave a dimension root on an ad hoc grid. The root Period member on a grid can also change data without setting impact status, so the habit costs integrity, not just speed.
The four causes, and the same-day fixes— what is computing when you thought you were reading.
CauseThe fix
Dynamic seeded parents on the gridThe Data Source parents, total input and adjusted, total eliminations, total data source, are dynamic calculations per Oracle's dimension reference, computed fresh at every retrieve, and the aggregating members above the belt behave the same way. Rebuild the grid on stored level-zero members and sum in the sheet; the retrieve drops from computation to read.
To-date views computed at retrieveQuick on periodic, crawling on year to date is the view storage decision showing its bill at the desk of whoever opened the grid. The fix is a design setting, not a grid change, and it has its own page.
Dimension roots as placeholdersA root member left on a grid asks for the whole dimension's aggregation. Saved grids with real members end both the wait and the silent impact-status failure the root Period member is documented to cause.
The everything gridWhole entity tree down the rows, every account across the columns, refreshed hourly during close. Scope grids like queries, entities and accounts the reader will look at, missing rows suppressed, and split the monster into the three grids people were extracting from it anyway.

The recurring heavy retrieve, the one refreshed hourly by five people, is not a grid problem, it is a report living in the wrong system. In the owned warehouse, the star schema we recommend building beside FCCS on your own cloud, mapped here, that retrieve is a query that costs the cube nothing. Member storage behavior verified against Oracle's dimension reference, July 2026.

Close week grids measured in minutes?
We find the computation hiding in your worst grids, fix the four causes, and move the hourly monsters to the star where they belong.
Talk to us
Terms on this page
dynamic calculation
Stored nowhere, computed at retrieve. The wait, most of the time.
stored member
Written at consolidation, read at retrieve. The fast path.
the everything grid
The whole cube in one sheet. Split it into what people read.