PHP Opc Ua
Menu
Home Documentation About
Sign in
uanetstandard-test-suite · master
Docs · Data features

Arrays and matrices

Arrays (1D), empty arrays, and 2D/3D matrices. Every shape your client's array handling needs to survive.

Path: TestServer / DataTypes / Array and MultiDimensional.

1D arrays (read-write)

Path: TestServer / DataTypes / Array

20 arrays with valueRank = 1. All RW.

Initial values come straight from src/TestServer/AddressSpace/DataTypesBuilder.cs. Most arrays are intentionally short (3 elements) — the suite optimises for boundary coverage and round-trip checks, not for stress data.

BrowseName ElementType Initial value
BooleanArray Boolean[] [true, false, true]
SByteArray SByte[] [-1, 0, 1]
ByteArray Byte[] [1, 2, 3]
Int16Array Int16[] [-100, 0, 100]
UInt16Array UInt16[] [100, 200, 300]
Int32Array Int32[] [-1000, 0, 1000]
UInt32Array UInt32[] [1000, 2000, 3000]
Int64Array Int64[] [-100000, 0, 100000]
UInt64Array UInt64[] [100000, 200000, 300000]
FloatArray Float[] [1.1, 2.2, 3.3]
DoubleArray Double[] [1.11, 2.22, 3.33]
StringArray String[] ["one", "two", "three"]
DateTimeArray DateTime[] [now, now + 1h, now + 2h] (captured at server startup)
GuidArray Guid[] Three fresh Guid.NewGuid() values per server start
ByteStringArray ByteString[] [ [0x01], [0x02], [0x03] ] (three single-byte byte strings)
XmlElementArray XmlElement[] Array.Empty<XmlElement>() (length 0)
NodeIdArray NodeId[] [ns=0;i=1, ns=0;i=2]
StatusCodeArray StatusCode[] [Good, Bad]
QualifiedNameArray QualifiedName[] [1:a, 1:b]
LocalizedTextArray LocalizedText[] [en:one, en:two]

1D arrays (read-only)

Path: TestServer / DataTypes / Array / ReadOnly

Six arrays with accessLevel = CurrentRead. Writes return Bad_NotWritable.

BrowseName Value
BooleanArray_RO [true, false]
Int32Array_RO [10, 20, 30]
DoubleArray_RO [1.0, 2.0, 3.0]
StringArray_RO ["read", "only"] (2 elements)
ByteArray_RO [0xAA, 0xBB] (2 elements)
DateTimeArray_RO [DateTime.UtcNow] (1 element, captured at startup)

Empty arrays

Path: TestServer / DataTypes / Array / Empty

14 variables initialised to []. Read-write. Used to verify your library handles zero-length arrays correctly — many encoders surprise themselves at this boundary.

BrowseName ElementType
EmptyBooleanArray Boolean[]
EmptySByteArray SByte[]
EmptyByteArray Byte[]
EmptyInt16Array Int16[]
EmptyUInt16Array UInt16[]
EmptyInt32Array Int32[]
EmptyUInt32Array UInt32[]
EmptyInt64Array Int64[]
EmptyUInt64Array UInt64[]
EmptyFloatArray Float[]
EmptyDoubleArray Double[]
EmptyStringArray String[]
EmptyDateTimeArray DateTime[]
EmptyByteStringArray ByteString[]

Read should return an array of length 0 (not null).

Multi-dimensional matrices

Path: TestServer / DataTypes / MultiDimensional

Three matrices with valueRank > 1. Stored as a flat array with arrayDimensions metadata, row-major ordering.

BrowseName Type Dimensions Initial value (row-major)
Matrix2D_Double Double 3 × 3 [[1,2,3], [4,5,6], [7,8,9]]
Matrix2D_Int32 Int32 2 × 4 [[1,2,3,4], [5,6,7,8]]
Cube3D_Byte Byte 2 × 3 × 4 new byte[2,3,4] — all 24 elements default to 0

To address element [i,j] of a 2D matrix in flat storage:

flat_index = i * cols + j

For 3D [i,j,k] with dimensions [dim0, dim1, dim2]:

flat_index = i * (dim1 * dim2) + j * dim2 + k

Reading the variable returns:

  • value: flat array
  • arrayDimensions: the shape ([3,3], [2,4], or [2,3,4])
  • valueRank: 2 or 3

Your client should expose both pieces — the flat data and the dimensions — so callers can reconstruct the matrix.

Test patterns

Round-trip 1D

For each RW 1D array:

write(<node>, <new-array>)
assert read(<node>) == <new-array>

Try different lengths than the initial value to exercise allocation behaviour.

Boundary values

Array Boundary check
SByteArray [Int8.MIN, Int8.MAX]
Int16Array [Int16.MIN, Int16.MAX]
UInt32Array Includes 4294967295 (UInt32.MAX)
UInt64Array Try with Int64.MAX + 1

Empty round-trip

For each empty array:

write(<node>, []) → expect Good
read(<node>)      → expect []

Then:

write(<node>, [a, b, c]) → expect Good
read(<node>)              → [a, b, c]
write(<node>, [])         → expect Good
read(<node>)              → []

Exercises the encoder's array-length handling at both extremes.

Matrix shape

For each multi-D matrix:

  1. Read once → get flat data + dimensions.
  2. Reconstruct the matrix locally.
  3. Write the reconstructed flat array back.
  4. Read again — bytes should match.

For Matrix2D_Double, also try writing a wrong-length flat array (8 instead of 9 elements). The server doesn't enforce length consistency, so the write succeeds — but reading back shows the inconsistency. Your client should ideally validate locally before writing.

Last updated · 2026-05-25
Edit on GitHub Report issue

← Previous

Scalar types

Next →

Structures and extension objects

Documentation