The Split Records from Array processor allows you to create many records from a single record, by splitting out new records for each element in an input array.

Use Split Records from Array where data that should be represented in many records has been wrongly captured in a single record; that is, to normalize data that has been wrongly denormalized.

Often, the denormalized data that you need to split out needs to be pre-processed, before using this processor. For example, in the following Orders table, multiple orders (with multiple order numbers and product descriptions) have been wrongly entered into a single record, using free text fields:

In this case, the Order_Number and Product_Desc attributes both need simple pre-processing using the Make Array from String processor to create arrays using the / character as a separator. The arrays can then be input into Split Records from Array to split out the records as follows:

Note above that many array attributes may be input to this processor. In this case, the number of output records for each input record will correspond to the number of elements in the array attribute with the largest number of elements. Data from each attribute that is not input is simply copied to all of the output records created from each input record. For example, if you split the following record, inputting Title.array and FirstName.array:

The output records will be as follows:

Note that Title.array.normalized is Null for the last record, as there is no array element in Title.array that corresponds with the third element in FirstName.array.

The following table describes the configuration options:

The following table describes the statistics produced by the profiler: