Large Data Objects in Landmark
The Lawson database can now store large objects (LOBs) and large binary objects (BLOBs). The database can hold fields of up to 4 GB in size, when the data type is specified to be a BLOB or CLOB. The field size only takes up as much actual space as is used by the object.
BLOB data processing within Lawson is optimized for large movement of data. Because the BLOB data is retrieved by a different mechanism, the overhead of LOBS is eliminated with standard data retrieval.
For help in troubleshooting issues with LOBs, see the dblobsizes utility.
The following large object types are supported:
| Lawson data type | Database data type | Description |
|---|---|---|
| Binary Document | BLOB |
This is a strongly-typed field used to store documents that are candidates for text indexing and searching (if the text search feature of the database is installed and enabled). Many formats are supported, including Word documents, RTF, PDF, HTML, or XML. The data in this field is stored as true binary. |
| Binary Image | BLOB | This is the same field type as a binary document (BLOB), but this field cannot be text indexed, and therefore text searching cannot be enabled for this field. This could be a graphics, video, or audio file; for example, file formats of mpeg, jpeg, or wav. |
| Text | CLOB | This is a large alpha field that is stored in the character set of the database. It can be indexed. |
lobspacename Considerations
Lawson recommends that BLOB data is stored out-of-line in a
separate tablespace, using the lobspacename attribute
of dbspace in the
dataarea.xml file. This allows more rows per block during record
retrieval of normal data – when reading in data you only get
the physical data. It also prevents rotating (minimizes row chaining)
which can occur when you add data to a parameter length field. This
expands the row size and throws more records out of memory.
If you don’t specify a lobspacename in
the
dataarea.xml file, then
BLOB data will be stored in-line within the table. This will significantly
affect the performance of data retrieval.