How chunking works during model write?

For Model write – default chunk size is 50 and developer can override that using chunkSize property in ModelWriteRequest

How chunking works during model write:

• Platform first do the chunking based on chunk size and not just based on top level but including all the children. For example, if the chunk size if 50 and if model list is having 1 shipment with 100 shipment lines, this will still be in a single chunk as platform doesn't check for chunk boundary for childrens.
• Then platform further divide the transaction list into sub blocks by separating the duplicate transactions
• And finally divide the transaction list into sub blocks grouped by action name

Apart from above and as part of 9.2, platform divide the transaction list into chunks if model has self-referential model link (but this happens after workflow invocation but before write).

For more information on self-referential, please check http://answers.elogex.com/questions/1195/self-referential-model-link-support