An opaque pointer to a collection schema object.

A PDCollectionSchema is ordered with the following considerations. The PDCollectionSchema has characteristics of both a collection of named fields and an array. Field names are guaranteed to be unique. Indexing is guaranteed to be 0-based, in ascending order, with no gaps. When repositioning a field by changing its index, the schema will be re-indexed as necessary to conform to this convention. If a field is assigned an out-of-bound index, the index is adjusted to be in bounds rather than raising an error.

The name of a field in a collection sort dictionary, and its associated value for ascending.

Acquires the PDCollectionSchema object for a collection.

Destroys a PDCollectionSchema object.

Gets a field by name or position. The caller must set field.size to sizeof(PDCollectionFieldRec). To look up a field by name, set field.fieldName to the appropriate name. To look up a field by position, set field.fieldName to ASAtomNull, and set field.index to the position. The caller owns (and must destroy) field.fieldText if it is not NULL.

Gets the number of fields in the schema.

Removes a field from the collection schema.

Sets a field with new values. The target field is identified by the field.fieldName member. If the target field exists, it is overwritten; otherwise a new field is added. The caller must set field.size to sizeof(PDCollectionFieldRec). Specifying a new value for field.index will affect other field values as necessary to maintain the correct overall ordering. See PDCollectionSchema for information about ordering.