Zoho Sheet DMIN Function: Criteria-Based Minimum

Zoho’s official DMIN documentation was retrieved on October 15, 2026. It documents the function in the rolling web service. The first release and last functioning version are unknown; this is not a historical build test. Mobile/offline parity is unverified. Results below are independently worked expectations, not captured Zoho Sheet execution. The datasets are fictional and fully specified.

DMIN(database; field; criteria)

All three arguments are documented. database includes a header row and its records. field identifies the column to process, using either its quoted header text or a reference to that header cell. criteria points to a separate range containing the relevant header and condition. The condition column and the result field can be different.

Use semicolons as shown in Zoho’s English reference; adjust separators if the workbook locale requires it. A header reference such as E1 is unquoted, whereas a literal name such as “Amount” needs quotes. These examples use explicit field names, not an assumed numeric field index or omitted argument.

Create a sheet named Jobs. Enter the following data in A1:F7. Headers are literal text; Units, Rate and Duration are numeric. In E2 enter =C2*D2 and fill through E7 to generate Amount. The displayed amounts below are checks, not extra inputs.

Set H1 to Team and H2 to Red. Set J1 to Team and J2 to Blue. These are two separate criteria ranges, H1:H2 and J1:J2. Red selects rows 2,4,5,7; Blue selects rows 3,6. Use the formulas on Jobs so the unqualified references point to this dataset.

The field-reference form uses F1, which contains the header Duration. Passing the text “F1” would instead supply that literal string as a field name. H1:H2 selects the team, although the result comes from Duration or Amount.

In L1 enter Team and leave L2 physically blank. =DMIN(A1:F7;"Duration";L1:L2) gives 8 under Zoho’s documented all-records rule. The Red-only and all-records results happen to agree because the shortest duration belongs to Red. Agreement in one dataset does not prove the filter was ignored.

Duplicate Jobs and change only F2 from 8 to 18. The Red duration result becomes 12; Blue remains 20. Restore the original table in a second copy and change H2 from Red to Blue. The result becomes 20 without changing any duration.

A useful additional control is to alter a nonselected row. On the original table, change Blue’s F6 from 30 to 100. Red’s DMIN stays 8, while Blue’s becomes 20. This separates filtering errors from an unexpected extreme in the selected group.

Use the pair of database functions to calculate Red’s observed duration range:

=DMAX(A1:F7;"Duration";H1:H2)-DMIN(A1:F7;"Duration";H1:H2)

The result is 16, from 24−8. Blue’s range is 10, from 30−20. The range measures the distance between endpoints; it does not summarize how intermediate durations cluster and is sensitive to a single extreme observation.

For a Red duration of 12 in F4, the distance above the selected minimum is =F4-DMIN(A1:F7;"Duration";H1:H2), expected 4. Keep the compared record in the same unit and group as the reference extreme.

The minimum duration is associated with T01 in this dataset, but DMIN returns only 8. If two records share that duration, the numerical result alone cannot identify a unique ticket. Review the matching records separately; do not treat the aggregate as a lookup function that selects a first or last match.

In a separate scenario, replace Red’s F2,F4,F5,F7 with -8,-12,-16,-24. DMIN becomes -24. The smallest negative number is the most negative, not the one with smallest magnitude. These signed numbers are a mathematical control, not plausible negative elapsed durations. Do not take absolute values unless the measurement definition requires them.

The exact native result for a selection with no numeric observations is not established here. In particular, a displayed zero should not be accepted as evidence that a genuine zero-valued record exists. Inspect the selected population before interpreting an extreme.

The official DMIN page lists shared name, type and reference errors. It does not exhaustively specify no-match behavior, mixed-type coercion, wildcard rules, duplicate header handling or every exact error token. Do not import those details from another spreadsheet product simply because the function name matches.

Start with the small dataset above. Confirm that the database includes row 1, the criteria header matches the intended field and all references point to the intended sheet. Inspect underlying cell contents rather than display formatting. Keep criteria outside the source records so editing a condition does not alter the database. A correct formula can still answer the wrong question if its range omits newly appended rows.

Before transferring a workbook, compare these controls in the destination app. This guide does not establish Excel, Numbers or other products’ syntax, error or version equivalence.

DMAX supplies the opposite endpoint. DAVERAGE describes the selected center; DCOUNT reports numeric participation.