SORTBY Function (LibreOffice Calc)

• Introduced: LibreOffice 24.8, explicitly stated in function-specific Help.
• Supported in: 26.8 Help and 26.8.1.1 registration; no retirement is established.
• Scope: Desktop Calc, English semicolon-separated formulas. The documented COM.MICROSOFT.SORTBY namespace is outside OpenFormula 1.3.
• Checked: October 14, 2026. Ten listed examples were recalculated in LibreOfficeDev 26.8.0.0.alpha0 using headless XLSX import. Stable releases, 26.8.1.1 and GUI array entry were not executed.

SORTBY returns a reordered copy of a range or array, using one or more corresponding key ranges. The keys can be outside the returned columns. A column-shaped key sorts records vertically; a row-shaped key sorts columns horizontally. The original cells are not rearranged. Function Help.

SORTBY(Range; SortByRange1 [; SortOrder1 [; SortByRange2 [; SortOrder2 …]]])

Square brackets mark optional arguments; do not type them. Range is the data to return. Each SortByRange supplies one key per returned row or column. SortOrder is 1 for ascending or −1 for descending; the default is ascending. For multiple keys, supply explicit direction arguments so the key/direction pairs remain clear. Earlier keys have priority, and later keys break ties. The inspected callback accepts the two-argument single-key form even though the Help’s displayed syntax lists SortOrder1 without brackets.

Enter this table in H10:J13:

For horizontal sorting, enter A, B, C, D across H15:K15 and 10, 40, 30, 20 across H16:K16.

The native column is from LibreOfficeDev 26.8.0.0.alpha0, build 2c87e51eeaa2b413ff4ae097b2705eea1995d8e5, using headless XLSX import and forced recalculation. Decimal values are rounded to twelve places; quoted text excludes the surrounding quotation marks from the returned value.

=SORTBY(H10:J13;I10:I13;1;J10:J13;-1) sorts the group ascending, then the score descending within each group. The labels become B, D, C, A: group 1 comes before group 2, and the larger score comes first inside each group. Sorting only by descending score would instead produce B, C, D, A. A secondary key therefore changes the meaning of the ranking rather than simply refining its appearance.

A key does not have to be included in the returned data. The same four scores can order the label-only range H10:H13 while keeping the score column out of the result. Shape must still agree: a vertical four-entry key describes four data rows, not four arbitrary cells. In the horizontal example, the row of scores orders the columns, giving labels A, D, C, B for ascending scores. SORTBY preserves the full returned width or height; it does not discard fields that were not used as keys. Inspected callback.

A key must be one-dimensional and match the relevant data dimension. The source checks these requirements and rejects direction values outside 1 and −1 after integer conversion. Invalid sizes and unusual fractional directions were not executed. No header is detected automatically: exclude header cells from both data and key ranges. The examples use numeric keys and distinct complete key combinations. Locale-dependent text collation and the order of rows tied on every key are not certified here.

SORT, INDEX, XMATCH.

• Function-specific Help
• Complete versioned callback
• English native-name registration

The saved formula identities, local input values and typed results were reconciled. The stale SUM cache changed to 3 and the uncached SUM returned 7. Saved formulas use COM.MICROSOFT.SORTBY for this operation. The examples establish only the listed development-build results; unexecuted guards and current-source behavior remain separately scoped.