#Analytics100 discussing

Hi @Ramkrishna Shirali - This is a schema mismatch error — the Event Monitoring data file being uploaded has 35 fields, but the dataset schema (defined in the dataflow's digest node) only expects 34 columns. The counts are off by one.

What caused it: Salesforce periodically adds new fields to Event Monitoring log files as part of seasonal releases. When they add a field to the LoginEvent log, the uploaded file gains a column but the CRM Analytics dataflow schema doesn't automatically update to match — hence the field-count vs header-count mismatch.

How to fix it:

1. Download a recent LoginEvent log file from Event Monitoring and open it. Identify which field is new (compare against the 34 columns your digest node expects — the extra one will stand out).
2. Update the digest node schema in the Login Upload flow dataflow to include the new field. You can do this in the dataflow JSON directly or via the UI dataflow editor — add the missing column with the appropriate data type.
3. Re-run the dataflow after saving the schema update.

Hi @Joost Breggeman,

This is a limitation — default filters in CRMA work well with 

standard user fields (like Name), not custom ones like Country/Region.

Make sure the field is in the dataset, but for custom fields you’ll likely need bindings or a dataset join. 

If you find this helpful feel free to mark it as Best Answer !!!! 😊

Thanks and Regards.

-- ── DATASET A ──────────────────────────────────────────

q_a = load "DatasetA";

q_b = load "DatasetB";

-- Step 1A: Get max amount per key from Dataset A

a_grouped = group q_a by DimensionKey;

a_max = foreach a_grouped generate

DimensionKey as DimensionKey,

max(q_a.Amount) as MaxAmount_A;

-- Step 2A: Join back to filter only rows where Amount = MaxAmount

a_joined = cogroup q_a by DimensionKey, a_max by DimensionKey;

a_at_max = foreach a_joined generate

q_a.DimensionKey as DimensionKey,

q_a.RevenueType as RevenueType,

q_a.Amount as Amount,

a_max.MaxAmount_A as MaxAmount_A;

a_at_max = filter a_at_max by Amount == MaxAmount_A;

-- Step 3A: Now first() is safe — all remaining rows have the max amount

a_grp = group a_at_max by DimensionKey;

a_final = foreach a_grp generate

DimensionKey,

first(a_at_max.RevenueType) as RevenueType_A;

-- ── DATASET B ──────────────────────────────────────────

-- Same pattern for Dataset B

b_grouped = group q_b by DimensionKey;

b_max = foreach b_grouped generate

DimensionKey as DimensionKey,

max(q_b.Amount) as MaxAmount_B;

b_joined = cogroup q_b by DimensionKey, b_max by DimensionKey;

b_at_max = foreach b_joined generate

q_b.DimensionKey as DimensionKey,

q_b.RevenueType as RevenueType,

q_b.Amount as Amount,

b_max.MaxAmount_B as MaxAmount_B;

b_at_max = filter b_at_max by Amount == MaxAmount_B;

b_grp = group b_at_max by DimensionKey;

b_final = foreach b_grp generate

DimensionKey,

first(b_at_max.RevenueType) as RevenueType_B;

-- ── COMBINE: Prefer A, fallback to B ───────────────────

combined = cogroup a_final by DimensionKey full,

b_final by DimensionKey full;

result = foreach combined generate

coalesce(a_final.DimensionKey, b_final.DimensionKey) as DimensionKey,

coalesce(a_final.RevenueType_A, b_final.RevenueType_B) as FinalRevenueType;

Yeah that's the classic toggle, one way around it: keep the static toggle but handle the "nothing selected = show all" logic in your step's SAQL. Instead of a straight filter binding, wrap it in a condition, something like if the selection is empty, return all rows, otherwise filter by the selected value. That way the dashboard loads with everything visible and users just click one toggle to drill down.

The SAQL would look roughly like:

q = load "your_dataset";
q = filter q by ({{column(toggle_step.selection, ["H","M","L"]).isEmpty()}} || 'your_bucket_field' in {{column(toggle_step.selection, ["value"])}});

The isEmpty() check means no selection = no filter = all data shown. One click on H and it filters to just H, no need to deselect the others.