Errata exists for this version of the document.

The OPC UA Binary DataEncoding is a data format developed to meet the performance needs of OPC UA applications. This format is designed primarily for fast encoding and decoding, however, the size of the encoded data on the wire was also a consideration.

The OPC UA Binary DataEncoding relies on several primitive data types with clearly defined encoding rules that can be sequentially written to or read from a binary stream. A structure is encoded by sequentially writing the encoded form of each field. If a given field is also a structure, then the values of its fields are written sequentially before writing the next field in the containing structure. All fields shall be written to the stream even if they contain null values. The encodings for each primitive type specify how to encode either a null or a default value for the type.

The OPC UA Binary DataEncoding does not include any type or field name information because all OPC UA applications are expected to have advance knowledge of the services and structures that they support. An exception is an ExtensionObject which provides an identifier and a size for the Structured DataType structure it represents. This allows a decoder to skip over types that it does not recognize.

A Boolean value shall be encoded as a single byte where a value of 0 (zero) is false and any non-zero value is true.

Encoders shall use the value of 1 to indicate a true value; however, decoders shall treat any non-zero value as true.

All integer types shall be encoded as little endian values where the least significant byte appears first in the stream.

Figure 2 illustrates how value 1 000 000 000 (Hex: 3B9ACA00) should be encoded as a 32-bit integer in the stream.

Figure 2 – Encoding Integers in a binary stream

All floating-point values shall be encoded with the appropriate IEEE-754 binary representation which has three basic components: the sign, the exponent, and the fraction. The bit ranges assigned to each component depend on the width of the type. Table 4 lists the bit ranges for the supported floating point types.

Table 4 – Supported Floating Point Types

Name	Width (bits)	Fraction	Exponent	Sign
Float	32	0-22	23-30	31
Double	64	0-51	52-62	63

In addition, the order of bytes in the stream is significant. All floating point values shall be encoded with the least significant byte appearing first (i.e. little endian).

Figure 3 illustrates how the value −6,5 (Hex: C0D00000) should be encoded as a Float.

The floating-point type supports positive and negative infinity and not-a-number (NaN). The IEEE specification allows for multiple NaN variants; however, the encoders/decoders may not preserve the distinction. Encoders shall encode a NaN value as an IEEE quiet-NAN (000000000000F8FF) or (0000C0FF). Any unsupported types such as denormalized numbers shall also be encoded as an IEEE quiet-NAN. Any test for equality between NaN values always fails.

Figure 3 – Encoding Floating Points in a binary stream

All String values are encoded as a sequence of UTF-8 characters without a null terminator and preceded by the length in bytes.

The length in bytes is encoded as Int32. A value of −1 is used to indicate a ‘null’ string.

Figure 4 illustrates how the multilingual string “水Boy” should be encoded in a byte stream.

Figure 4 – Encoding Strings in a binary stream

A DateTime value shall be encoded as a 64-bit signed integer (see Clause 5.2.2.2) which represents the number of 100 nanosecond intervals since January 1, 1601 (UTC).

Not all DevelopmentPlatforms will be able to represent the full range of dates and times that can be represented with this DataEncoding. For example, the UNIX time_t structure only has a 1 second resolution and cannot represent dates prior to 1970. For this reason, a number of rules shall be applied when dealing with date/time values that exceed the dynamic range of a DevelopmentPlatform. These rules are:

1. A date/time value is encoded as 0 if either
2. The value is equal to or earlier than 1601-01-01 12:00AM UTC.
3. The value is the earliest date that can be represented with the DevelopmentPlatform’s encoding.
4. A date/time is encoded as the maximum value for an Int64 if either
5. The value is equal to or greater than 9999-12-31 11:59:59PM UTC,
6. The value is the latest date that can be represented with the DevelopmentPlatform’s encoding.
7. A date/time is decoded as the earliest time that can be represented on the platform if either
8. The encoded value is 0,
9. The encoded value represents a time earlier than the earliest time that can be represented with the DevelopmentPlatform’s encoding.
10. A date/time is decoded as the latest time that can be represented on the platform if either
11. The encoded value is the maximum value for an Int64,
12. The encoded value represents a time later than the latest time that can be represented with the DevelopmentPlatform’s encoding.

These rules imply that the earliest and latest times that can be represented on a given platform are invalid date/time values and should be treated that way by applications.

A decoder shall truncate the value if a decoder encounters a DateTime value with a resolution that is greater than the resolution supported on the DevelopmentPlatform.

A Guid is encoded in a structure as shown in Table 2. Fields are encoded sequentially according to the data type for field.

Figure 5 illustrates how the Guid “72962B91-FA75-4AE6-8D28-B404DC7DAF63” should be encoded in a byte stream.

Figure 5 – Encoding Guids in a binary stream

A ByteString is encoded as sequence of bytes preceded by its length in bytes. The length is encoded as a 32-bit signed integer as described above.

If the length of the byte string is −1 then the byte string is ‘null’.

An XmlElement is an XML fragment serialized as UTF-8 string and then encoded as ByteString.

Figure 6 illustrates how the XmlElement “<A>Hot水</A>” should be encoded in a byte stream.

.

Figure 6 – Encoding XmlElement in a binary stream

A decoder may choose to parse the XML after decoding; if an unrecoverable parsing error occurs then the decoder should try to continue processing the stream. For example, if the XmlElement is the body of a Variant or an element in an array which is the body of a Variant then this error can be reported by setting value of the Variant to the StatusCode Bad_DecodingError.

The components of a NodeId are described the Table 5.

Table 5 – NodeId components

Name	Data Type	Description
Namespace	UInt16	The index for a namespace URI. An index of 0 is used for OPC UA defined NodeIds.
IdentifierType	Enumeration	The format and data type of the identifier. The value may be one of the following: NUMERIC - the value is an UInteger; STRING - the value is String; GUID - the value is a Guid; OPAQUE- the value is a ByteString;
Value	*	The identifier for a node in the address space of an OPC UA Server.

The DataEncoding of a NodeId varies according to the contents of the instance. For that reason, the first byte of the encoded form indicates the format of the rest of the encoded NodeId. The possible DataEncoding formats are shown in Table 6. The tables that follow describe the structure of each possible format (they exclude the byte which indicates the format).

Table 6 – NodeId DataEncoding values

Name	Value	Description
Two Byte	0x00	A numeric value that fits into the two-byte representation.
Four Byte	0x01	A numeric value that fits into the four-byte representation.
Numeric	0x02	A numeric value that does not fit into the two or four byte representations.
String	0x03	A String value.
Guid	0x04	A Guid value.
ByteString	0x05	An opaque (ByteString) value.
NamespaceUri Flag	0x80	See discussion of ExpandedNodeId in 5.2.2.10.
ServerIndex Flag	0x40	See discussion of ExpandedNodeId in 5.2.2.10.

The standard NodeId DataEncoding has the structure shown in Table 7. The standard DataEncoding is used for all formats that do not have an explicit format defined.

Table 7 – Standard NodeId Binary DataEncoding

Name	Data Type	Description
Namespace	UInt16	The NamespaceIndex.
Identifier	*	The identifier which is encoded according to the following rules: NUMERIC UInt32 STRING String GUID Guid OPAQUE ByteString

An example of a String NodeId with Namespace = 1 and Identifier = “Hot水” is shown in Figure 7.

Figure 7 – A String NodeId

The Two Byte NodeId DataEncoding has the structure shown in Table 8.

Table 8 – Two Byte NodeId Binary DataEncoding

Name	Data Type	Description
Identifier	Byte	The Namespace is the default OPC UA namespace (i.e. 0). The Identifier Type is ‘Numeric’. The Identifier shall be in the range 0 to 255.

An example of a Two Byte NodeId with Identifier = 72 is shown in Figure 8.

Figure 8 – A Two Byte NodeId

The Four Byte NodeId DataEncoding has the structure shown in Table 9.

Table 9 – Four Byte NodeId Binary DataEncoding

Name	Data Type	Description
Namespace	Byte	The Namespace shall be in the range 0 to 255.
Identifier	UInt16	The Identifier Type is ‘Numeric’. The Identifier shall be an integer in the range 0 to 65 535.

An example of a Four Byte NodeId with Namespace = 5 and Identifier = 1 025 is shown in Figure 9.

Figure 9 – A Four Byte NodeId

An ExpandedNodeId extends the NodeId structure by allowing the NamespaceUri to be explicitly specified instead of using the NamespaceIndex. The NamespaceUri is optional. If it is specified, then the NamespaceIndex inside the NodeId shall be ignored.

The ExpandedNodeId is encoded by first encoding a NodeId as described in 5.2.2.9 and then encoding NamespaceUri as a String.

An instance of an ExpandedNodeId may still use the NamespaceIndex instead of the NamespaceUri. In this case, the NamespaceUri is not encoded in the stream. The presence of the NamespaceUri in the stream is indicated by setting the NamespaceUri flag in the encoding format byte for the NodeId.

If the NamespaceUri is present, then the encoder shall encode the NamespaceIndex as 0 in the stream when the NodeId portion is encoded. The unused NamespaceIndex is included in the stream for consistency.

An ExpandedNodeId may also have a ServerIndex which is encoded as a UInt32 after the NamespaceUri. The ServerIndex flag in the NodeId encoding byte indicates whether the ServerIndex is present in the stream. The ServerIndex is omitted if it is equal to zero.

The ExpandedNodeId encoding has the structure shown in Table 10.

Table 10 – ExpandedNodeId Binary DataEncoding

Name	Data Type	Description
NodeId	NodeId	The NamespaceUri and ServerIndex flags in the NodeId encoding indicate whether those fields are present in the stream.
NamespaceUri	String	Not present if null or Empty.
ServerIndex	UInt32	Not present if 0.

A StatusCode is encoded as a UInt32.

A DiagnosticInfo structure is described in OPC 10000-4. It specifies a number of fields that could be missing. For that reason, the encoding uses a bit mask to indicate which fields are actually present in the encoded form.

As described in OPC 10000-4, the SymbolicId, NamespaceUri, LocalizedText and Locale fields are indexes in a string table which is returned in the response header. Only the index of the corresponding string in the string table is encoded. An index of −1 indicates that there is no value for the string.

DiagnosticInfo allows unlimited nesting which could result in stack overflow errors even if the message size is less than the maximum allowed. Decoders shall support at least 100 nesting levels. Decoders shall report an error if the number of nesting levels exceeds what it supports.

Table 11 – DiagnosticInfo Binary DataEncoding

Name	Data Type	Description
Encoding Mask	Byte	A bit mask that indicates which fields are present in the stream. The mask has the following bits: 0x01 Symbolic Id 0x02 Namespace 0x04 LocalizedText 0x08 Locale 0x10 Additional Info 0x20 InnerStatusCode 0x40 InnerDiagnosticInfo
SymbolicId	Int32	A symbolic name for the status code.
NamespaceUri	Int32	A namespace that qualifies the symbolic id.
Locale	Int32	The locale used for the localized text.
LocalizedText	Int32	A human readable summary of the status code.
Additional Info	String	Detailed application specific diagnostic information.
Inner StatusCode	StatusCode	A status code provided by an underlying system.
Inner DiagnosticInfo	DiagnosticInfo	Diagnostic info associated with the inner status code.

A QualifiedName structure is encoded as shown in Table 12.

The abstract QualifiedName structure is defined in OPC 10000-3.

Table 12 – QualifiedName Binary DataEncoding

Name	Data Type	Description
NamespaceIndex	UInt16	The namespace index.
Name	String	The name.

A LocalizedText structure contains two fields that could be missing. For that reason, the encoding uses a bit mask to indicate which fields are actually present in the encoded form.

The abstract LocalizedText structure is defined in OPC 10000-3.

Table 13 – LocalizedText Binary DataEncoding

Name	Data Type	Description
EncodingMask	Byte	A bit mask that indicates which fields are present in the stream. The mask has the following bits: 0x01 Locale 0x02 Text
Locale	String	The locale. Omitted is null or empty.
Text	String	The text in the specified locale. Omitted is null or empty.

An ExtensionObject is encoded as sequence of bytes prefixed by the NodeId of its DataTypeEncoding and the number of bytes encoded.

An ExtensionObject may be encoded by the application which means it is passed as a ByteString or an XmlElement to the encoder. In this case, the encoder will be able to write the number of bytes in the object before it encodes the bytes. However, an ExtensionObject may know how to encode/decode itself which means the encoder shall calculate the number of bytes before it encodes the object or it shall be able to seek backwards in the stream and update the length after encoding the body.

When a decoder encounters an ExtensionObject it shall check if it recognizes the DataTypeEncoding identifier. If it does, then it can call the appropriate function to decode the object body. If the decoder does not recognize the type it shall use the Encoding to determine if the body is a ByteString or an XmlElement and then decode the object body or treat it as opaque data and skip over it.

The serialized form of an ExtensionObject is shown in Table 14.

Table 14 – Extension Object Binary DataEncoding

Name	Data Type	Description
TypeId	NodeId	The identifier for the DataTypeEncoding node in the Server's AddressSpace. ExtensionObjects defined by the OPC UA specification have a numeric node identifier assigned to them with a NamespaceIndex of 0. The numeric identifiers are defined in A.3. Decoders use this field to determine the syntax of the Body. For example, if this field is the NodeId of the JSON Encoding Object for a DataType then the Body is a ByteString containing a JSON document encoded as a UTF-8 string.
Encoding	Byte	An enumeration that indicates how the body is encoded. The parameter may have the following values: 0x00 No body is encoded. 0x01 The body is encoded as a ByteString. 0x02 The body is encoded as a XmlElement.
Length	Int32	The length of the object body. The length shall be specified if the body is encoded.
Body	Byte [*]	The object body. This field contains the raw bytes for ByteString bodies. For XmlElement bodies this field contains the XML encoded as a UTF-8 string without any null terminator. Some binary encoded structures may have a serialized length that is not a multiple of 8 bits. Encoders shall append 0 bits to ensure the serialized length is a multiple of 8 bits. Decoders that understand the serialized format shall ignore the padding bits.

ExtensionObjects are used in two contexts: as values contained in Variant structures or as parameters in OPC UA Messages.

A decoder may choose to parse an XmlElement body after decoding; if an unrecoverable parsing error occurs then the decoder should try to continue processing the stream. For example, if the ExtensionObject is the body of a Variant or an element in an array that is the body of Variant then this error can be reported by setting value of the Variant to the StatusCode Bad_DecodingError.

A Variant is a union of the built-in types.

The structure of a Variant is shown in Table 15.

Table 15 – Variant Binary DataEncoding

Name	Data Type	Description
EncodingMask	Byte	The type of data encoded in the stream. A value of 0 specifies a NULL and that no other fields are encoded.The mask has the following bits assigned: 0:5 Built-in Type Id (see Table 1). 6 True if the Array Dimensions field is encoded. 7 True if an array of values is encoded. The Built-in Type Ids 26 through 31 are not currently assigned but may be used in the future. Decoders shall accept these IDs, assume the Value contains a ByteString and pass both onto the application. Encoders shall not use these IDs.
ArrayLength	Int32	The number of elements in the array. This field is only present if the array bit is set in the encoding mask. Multi-dimensional arrays are encoded as a one-dimensional array and this field specifies the total number of elements. The original array can be reconstructed from the dimensions that are encoded after the value field. Higher rank dimensions are serialized first. For example, an array with dimensions [2,2,2] is written in this order: [0,0,0], [0,0,1], [0,1,0], [0,1,1], [1,0,0], [1,0,1], [1,1,0], [1,1,1]
Value	*	The value encoded according to its built-in data type. If the array bit is set in the encoding mask, then each element in the array is encoded sequentially. Since many types have variable length encoding each element shall be decoded in order. The value shall not be a Variant but it could be an array of Variants. Many implementation platforms do not distinguish between one dimensional Arrays of Bytes and ByteStrings. For this reason, decoders are allowed to automatically convert an Array of Bytes to a ByteString.
ArrayDimensionsLength	Int32	The number of dimensions. This field is only present if the ArrayDimensions flag is set in the encoding mask.
ArrayDimensions	Int32[*]	The length of each dimension encoded as a sequence of Int32 values This field is only present if the ArrayDimensions flag is set in the encoding mask. The lower rank dimensions appear first in the array. All dimensions shall be specified and shall be greater than zero. If ArrayDimensions are inconsistent with the ArrayLength then the decoder shall stop and raise a Bad_DecodingError.

The types and their identifiers that can be encoded in a Variant are shown in Table 1.

A DataValue is always preceded by a mask that indicates which fields are present in the stream.

The fields of a DataValue are described in Table 16.

Table 16 – Data Value Binary DataEncoding

Name	Data Type	Description
Encoding Mask	Byte	A bit mask that indicates which fields are present in the stream. The mask has the following bits: 0x01 False if the Value is Null. 0x02 False if the StatusCode is Good. 0x04 False if the Source Timestamp is DateTime.MinValue. 0x08 False if the Server Timestamp is DateTime.MinValue. 0x10 False if the Source Picoseconds is 0. 0x20 False if the Server Picoseconds is 0.
Value	Variant	The value. Not present if the Value bit in the EncodingMask is False.
Status	StatusCode	The status associated with the value. Not present if the StatusCode bit in the EncodingMask is False.
SourceTimestamp	DateTime	The source timestamp associated with the value. Not present if the SourceTimestamp bit in the EncodingMask is False.
SourcePicoSeconds	UInt16	The number of 10 picosecond intervals for the SourceTimestamp. Not present if the SourcePicoSeconds bit in the EncodingMask is False. If the source timestamp is missing the picoseconds are ignored.
ServerTimestamp	DateTime	The Server timestamp associated with the value. Not present if the ServerTimestamp bit in the EncodingMask is False.
ServerPicoSeconds	UInt16	The number of 10 picosecond intervals for the ServerTimestamp. Not present if the ServerPicoSeconds bit in the EncodingMask is False. If the Server timestamp is missing the picoseconds are ignored.

The Picoseconds fields store the difference between a high-resolution timestamp with a resolution of 10 picoseconds and the Timestamp field value which only has a 100 ns resolution. The Picoseconds fields shall contain values less than 10 000. The decoder shall treat values greater than or equal to 10 000 as the value ‘9999’.

Decimals are encoded as described in 5.1.7.

A Decimal does not have a NULL value.

Enumerations are encoded as Int32 values.

An Enumeration does not have a NULL value.

One dimensional Arrays are encoded as a sequence of elements preceded by the number of elements encoded as an Int32 value. If an Array is null, then its length is encoded as −1. An Array of zero length is different from an Array that is null so encoders and decoders shall preserve this distinction.

Multi-dimensional Arrays are encoded as an Int32 Array containing the dimensions followed by a list of all the values in the Array. The total number of values is equal to the product of the dimensions. The number of values is 0 if one or more dimension is less than or equal to 0. The process for reconstructing the multi-dimensional array is described in 5.2.2.16.

Structures are encoded as a sequence of fields in the order that they appear in the definition. The encoding for each field is determined by the built-in type for the field.

All fields specified in the structure shall be encoded. If optional fields exist in the structure then see 5.2.7.

Structures do not have a null value. If an encoder is written in a programming language that allows structures to have null values, then the encoder shall create a new instance with default values for all fields and serialize that. Encoders shall not generate an encoding error in this situation.

The following is an example of a structure using C/C++ syntax:

struct Type2

{

Int32 A;

Int32 B;

};

struct Type1

{

Int32 X;

Byte NoOfY;

Type2* Y;

Int32 Z;

};

In the C/C++ example above, the Y field is a pointer to an array with a length stored in NoOfY. When encoding an array the length is part of the array encoding so the NoOfY field is not encoded. That said, encoders and decoders use NoOfY during encoding.

An instance of Type1 which contains an array of two Type2 instances would be encoded as 28-byte sequence. If the instance of Type1 was encoded in an ExtensionObject it would have an additional prefix shown in Table 17 which would make the total length 37 bytes The TypeId, Encoding and the Length are fields defined by the ExtensionObject. The encoding of the Type2 instances do not include any type identifier because it is explicitly defined in Type1.

Table 17 – Sample OPC UA Binary Encoded structure

Field	Bytes	Value
Type Id	4	The identifier for Type1
Encoding	1	0x1 for ByteString
Length	4	28
X	4	The value of field ‘X’
Y.Length	4	2
Y.A	4	The value of field ‘Y[0].A’
Y.B	4	The value of field ‘Y[0].B’
Y.A	4	The value of field ‘Y[1].A’
Y.B	4	The value of field ‘Y[1].B’
Z	4	The value of field ‘Z’

The Value of the DataTypeDefinition Attribute for a DataType Node describing Type1 is:

Name	Type	Description
defaultEncodingId	NodeId	NodeId of the “Type1_Encoding_DefaultBinary” Node.
baseDataType	NodeId	“i=22” [Structure]
structureType	StructureType	Structure_0 [Structure without optional fields]
fields [0]	StructureField
name	String	“X”
description	LocalizedText	Description of X
dataType	NodeId	“i=6” [Int32]
valueRank	Int32	-1 (Scalar)
isOptional	Boolean	false
fields [1]	StructureField
name	String	“Y“
description	LocalizedText	Description of Y-Array
dataType	NodeId	NodeId of the Type2 DataType Node (e.g. “ns=3; s=MyType2”)
valueRank	Int32	1 (OneDimension)
isOptional	Boolean	false
fields [2]	StructureField
name	String	“Z“
description	LocalizedText	Description of Z
dataType	NodeId	“i=6” [Int32]
valueRank	Int32	-1 (Scalar)
isOptional	Boolean	false

The Value of the DataTypeDefinition Attribute for a DataType Node describing Type2 is:

Name	Type	Description
defaultEncodingId	NodeId	NodeId of the “Type2_Encoding_DefaultBinary” Node.
baseDataType	NodeId	“i=22” [Structure]
structureType	StructureType	Structure_0 [Structure without optional fields]
fields [0]	StructureField
name	String	“A“
description	LocalizedText	Description of A
dataType	NodeId	“i=6” [Int32]
valueRank	Int32	-1 (Scalar)
isOptional	Boolean	false
fields [1]	StructureField
name	String	“B“
description	LocalizedText	Description of B
dataType	NodeId	“i=6” [Int32]
valueRank	Int32	-1 (Scalar)
isOptional	Boolean	false

Structures with optional fields are encoded with an encoding mask preceding a sequence of fields in the order that they appear in the definition. The encoding for each field is determined by the data type for the field.

The EncodingMask is a 32-bit unsigned integer. Each optional field is assigned exactly one bit. The first optional field is assigned bit ‘0’, the second optional field is assigned bit ‘1’ and so until all optional fields are assigned bits. A maximum of 32 optional fields can appear within a single Structure. Unassigned bits are set to 0 by encoders. Decoders shall report an error if unassigned bits are not 0.

The following is an example of a structure with optional fields using C++ syntax:

struct TypeA

{

Int32 X;

Int32* O1;

SByte Y;

Int32* O2;

};

O1 and O2 are optional fields which are NULL if not present

An instance of TypeA which contains two mandatory (X and Y) and two optional (O1 and O2) fields would be encoded as a byte sequence. The length of the byte sequence is depending on the available optional fields. An encoding mask field determines the available optional fields.

An instance of TypeA where field O2 is available and field O1 is not available would be encoded as a 13-byte sequence. If the instance of TypeA was encoded in an ExtensionObject it would have the encoded form shown in Table 17 and have a total length of 20 bytes. The length of the TypeId, Encoding and the Length are fields defined by the ExtensionObject.

Table 17 – Sample OPC UA Binary Encoded Structure with optional fields

Field	Bytes	Value
Type Id	4	The identifier for TypeA
Encoding	1	0x1 for ByteString
Length	4	13
EncodingMask	4	0x02 for O2
X	4	The value of X
Y	1	The value of Y
O2	4	The value of *O2

If a Structure with optional fields is subtyped, the subtypes extend the EncodingMask defined for the the parent.

The Value of the DataTypeDefinition Attribute for a DataType Node describing TypeA is:

Name	Type	Description
defaultEncodingId	NodeId	NodeId of the “TypeA_Encoding_DefaultBinary” Node.
baseDataType	NodeId	“i=22” [Structure]
structureType	StructureType	StructureWithOptionalFields_1 [Structure without optional fields]
fields [0]	StructureField
name	String	“X”
description	LocalizedText	Description of X
dataType	NodeId	“i=6” [Int32]
valueRank	Int32	-1 (Scalar)
isOptional	Boolean	false
fields [1]	StructureField
name	String	“O1“
description	LocalizedText	Description of O1
dataType	NodeId	“i=6” [Int32]
valueRank	Int32	-1 (Scalar)
isOptional	Boolean	true
fields [2]	StructureField
name	String	“Y“
description	LocalizedText	Description of Z
dataType	NodeId	“i=2” [SByte]
valueRank	Int32	-1 (Scalar)
isOptional	Boolean	false
fields [3]	StructureField
name	String	“O2“
description	LocalizedText	Description of O2
dataType	NodeId	“i=6” [Int32]
valueRank	Int32	-1 (Scalar)
isOptional	Boolean	true

Unions are encoded as a switch field preceding one of the possible fields. The encoding for the selected field is determined by the data type for the field.

The switch field is encoded as a UInt32.

The switch field is the index of the available union fields starting with 1. If the switch field is 0 then no field is present. For any value greater than the number of defined union fields the encoders or decoders shall report an error.

A Union with no fields present has the same meaning as a NULL value. A Union with any field present is not a NULL value even if the value of the field itself is NULL.

The following is an example of a union using C/C++ syntax:

struct Type2

{

Int32 A;

Int32 B;

};

struct Type1

{

Byte Selector;

union

{

Int32Field1;

Type2Field2;

}

Value;

};

In the C/C++ example above, the Selector and Value are semantically coupled to form a union. The order of the fields does not matter.

An instance of Type1 would be encoded as byte sequence. The length of the byte sequence depends on the selected field.

An instance of Type1 where field Field1 is available would be encoded as 8-byte sequence. If the instance of Type 1 was encoded in an ExtensionObject it would have the encoded form shown in Table 18 and it would have a total length of 15 bytes. The TypeId, Encoding and the Length are fields defined by the ExtensionObject.

Table 18 – Sample OPC UA Binary Encoded Structure

Field	Bytes	Value
Type Id	4	The identifier for Type1
Encoding	1	0x1 for ByteString
Length	4	8
SwitchValue	4	1 for Field1
Field1	4	The value of Field1

The Value of the DataTypeDefinition Attribute for a DataType Node describing Type1 is:

Name	Type	Description
defaultEncodingId	NodeId	NodeId of the “Type1_Encoding_DefaultBinary” Node.
baseDataType	NodeId	“i=22” [Union]
structureType	StructureType	Union_2 [Union]
fields [0]	StructureField
name	String	“Field1”
description	LocalizedText	Description of Field1
dataType	NodeId	“i=6” [Int32]
valueRank	Int32	-1 (Scalar)
isOptional	Boolean	true
fields [1]	StructureField
name	String	“Field2“
description	LocalizedText	Description of Field2
dataType	NodeId	NodeId of the Type2 DataType Node (e.g. “ns=3; s=MyType2”)
valueRank	Int32	-1 (Scalar)
isOptional	Boolean	true

The Value of the DataTypeDefinition Attribute for a DataType Node describing Type2 is:

Name	Type	Description
defaultEncodingId	NodeId	NodeId of the “Type2_Encoding_DefaultBinary” Node.
baseDataType	NodeId	“i=22” [Structure]
structureType	StructureType	Structure_0 [Structure without optional fields]
fields [0]	StructureField
name	String	“A“
description	LocalizedText	Description of A
dataType	NodeId	“i=6” [Int32]
valueRank	Int32	-1 (Scalar)
isOptional	Boolean	false
fields [1]	StructureField
name	String	“B“
description	LocalizedText	Description of B
dataType	NodeId	“i=6” [Int32]
valueRank	Int32	-1 (Scalar)
isOptional	Boolean	false

Messages are Structures encoded as sequence of bytes prefixed by the NodeId of for the OPC UA Binary DataTypeEncoding defined for the Message.

Each OPC UA Service described in OPC 10000-4 has a request and response Message. The DataTypeEncoding IDs assigned to each Service are specified in A.3.