Data Conversion:
** The 3rd method to convert the data is through Data conversion t/r directly.
**When we convert the data using type casting by derived/conditional split t/r's we can directly
perform any operations on converted data, but here when we use this t/r a new column is added
in the output.
** Then the above output can be used for further processing.
Example: Try with a flat file source as input.
Merge Transformation:
** Up to now we have seen only one source as input but we can have multiple sources by merge t/r.
** For merge t/r we can only have 2 i/p sources and that to of similar structures and same data types
** The i/p to the t/r should be in sorted order based on the same column in both the i/p's
Example: Take a flat file source and a oledb source, then convert the column which needs to be merged into the same data type at source level. Link the sources to sort t/r's and define the sort column( same column in both sources and in same order). Link the sort t/r's to merge t/r.
Note : We can directly provide the sorted data to merge t/r's without using sort t/r's but we need to specify the source is already sorted.
** for the above criteria select the source, right click show advanced editor, select i/p % o/p properties tab. Select the source in properties window set the property IsSorted to true.
** Expand the source, expand the o/p columns folder. Select the column which is already sorted, set the property sortkeyposition to 1( asec) for desc(-1).
Union All :
** This t/r merge's multiple sources except the data need not be sorted and o/p may not be sorted.
** Same as union all in SQL.
** The 3rd method to convert the data is through Data conversion t/r directly.
**When we convert the data using type casting by derived/conditional split t/r's we can directly
perform any operations on converted data, but here when we use this t/r a new column is added
in the output.
** Then the above output can be used for further processing.
Example: Try with a flat file source as input.
Merge Transformation:
** Up to now we have seen only one source as input but we can have multiple sources by merge t/r.
** For merge t/r we can only have 2 i/p sources and that to of similar structures and same data types
** The i/p to the t/r should be in sorted order based on the same column in both the i/p's
Example: Take a flat file source and a oledb source, then convert the column which needs to be merged into the same data type at source level. Link the sources to sort t/r's and define the sort column( same column in both sources and in same order). Link the sort t/r's to merge t/r.
Note : We can directly provide the sorted data to merge t/r's without using sort t/r's but we need to specify the source is already sorted.
** for the above criteria select the source, right click show advanced editor, select i/p % o/p properties tab. Select the source in properties window set the property IsSorted to true.
** Expand the source, expand the o/p columns folder. Select the column which is already sorted, set the property sortkeyposition to 1( asec) for desc(-1).
Union All :
** This t/r merge's multiple sources except the data need not be sorted and o/p may not be sorted.
** Same as union all in SQL.
Comments
Post a Comment