MMULT in Collabora Office: Multiply Matrices With Correct Dimensions
Calculate Collabora MMULT using complete rectangular matrices, row-by-column products and transparent weighted-total controls.
Verification status: See the evidence and test scope described in this guide. How formula examples are checked
Quick answer: MMULT performs matrix multiplication: each output entry is a row-by-column dot product. If A has r rows and k columns and B has k rows and c columns, the product has r rows and c columns.
Availability and version history
- Introduced: original Collabora release and launch date are not established by the sources reviewed.
- Availability: documented in Collabora Office 26.04 help, scoped to the desktop-help release.
- Platform scope: support in every online or mobile deployment is not inferred from desktop documentation.
- Checked: October 14, 2026. Function-specific documentation and independently derived controls; no native Collabora test was performed.
Syntax and the dimension rule
=MMULT(Array1;Array2)
Both arguments are required. The direct Collabora description says the first array’s column count must equal the second array’s row count. One later parameter sentence is imprecise about “the same number of rows”; use the explicit column-to-row requirement rather than inferring that the inputs must have identical shapes.
Matrix multiplication differs from multiplying corresponding cells. The examples use all-numeric ranges and explicitly compute every result entry.
Enter the complete array result
These examples follow the desktop array-entry workflow documented for Collabora Office 26.04. Select the full output rectangle before entering a multi-cell formula. In the Function Wizard, select Array; for direct entry use the client-appropriate array-entry shortcut shown in the help, Ctrl+Shift+Enter or Command+Shift+Enter. Do not type the displayed outer braces manually.
Input cells and output cells must be separate. The help describes a linked array result whose cells are edited as a group; selecting too small an output rectangle does not redefine the mathematics of the source arrays. This workflow is not a claim that every online or mobile client has the same controls.
Example 1: multiply a two-by-three by a three-by-two matrix
Enter the first matrix in A2:C3:
| Worksheet row | A | B | C |
|---|---|---|---|
| 2 | 2 | 0 | 1 |
| 3 | 1 | 3 | −1 |
Enter the second matrix in E2:F4:
| Worksheet row | E | F |
|---|---|---|
| 2 | 4 | 2 |
| 3 | 0 | 5 |
| 4 | 6 | −2 |
The inner dimensions are both three. Select H2:I3 for the two-by-two result and use array entry:
=MMULT(A2:C3;E2:F4)
Expected output:
| Worksheet row | H | I |
|---|---|---|
| 2 | 14 | 2 |
| 3 | −2 | 19 |
The four independent calculations are 2×4+0×0+1×6=14; 2×2+0×5+1×(−2)=2; 1×4+3×0+(−1)×6=−2; and 1×2+3×5+(−1)×(−2)=19.
Example 2: reverse the factors
The reversed product is also dimensionally valid, but its shape changes to three-by-three. Select a fresh three-by-three output rectangle and enter:
=MMULT(E2:F4;A2:C3)
Expected rows are (10,6,2), (5,15,−5) and (10,−6,8). This is not the same output as the previous two-by-two result. Matrix multiplication is generally order-dependent; validity in both orders does not imply equality.
Check the first reversed row: 4 times (2,0,1) plus 2 times (1,3,−1) gives (10,6,2). That reconstruction makes the dimensions and operation visible.
Advanced example: apply one weight vector
Retain A2:C3 and enter numeric weights 10,20,30 in K2:K4. Select M2:M3 and enter:
=MMULT(A2:C3;K2:K4)
Expected output is the two-row column 50,40. The first result is 2×10+0×20+1×30; the second is 1×10+3×20−1×30.
The negative coefficient is an explicit mathematical input, not an assumption about physical quantities. The weights must correspond in order to columns A, B and C. Rearranging one side alone changes the calculation.
Advanced check: identity multiplication
Create a three-by-three identity matrix using MUNIT in a separate J7:L9 block. Then select a two-by-three result area and enter:
=MMULT(A2:C3;J7:L9)
Expected output is the original matrix with rows (2,0,1) and (1,3,−1). Multiplication by an identity of the correct size preserves the matrix.
A two-by-two identity belongs on the left of this two-by-three matrix, not on its right. The inner-dimension rule determines that distinction before any arithmetic is performed.
Troubleshooting and unresolved errors
A2:C3 multiplied by another two-row matrix with only two rows fails the required inner-dimension match: three does not equal two. The cited Collabora section does not provide a complete error-code, mixed-type or maximum-size table, so this guide does not assign native codes to those invalid fixtures.
Do not infer a square-output restriction from the help’s square-range example. The general matrix-product shape rule is r-by-c, as the vector and reversed-product controls demonstrate mathematically. Native behavior for truncated output selections, arbitrary coercions and extreme precision remains untested.
See MINVERSE for inverse matrices, or TRANSPOSE to swap rows and columns when that is the intended transformation. Transposing an input merely to make dimensions fit changes its meaning and needs its own justification.
References and verification
Complete matrices and outputs were independently calculated using arithmetic and exact rational checks where appropriate. They are documentation-backed expectations, not native Collabora captures or a certification of every type, error and client boundary.