Data in Data Flows
Integration Services provides a set of data types that are used in data flows.
The source that you add to a data flow converts the source data to Integration Services data types. Subsequent transformations may convert the data to different Integration Services data types, and depending on the type of data store into which data is loaded, destinations may convert the final Integration Services data type to the data type required by the destination data store. For more information, see Integration Services Data Types.
To convert the data to an Integration Services data type, a data flow component parses the data. Integration Services provides two types of data parsing: fast parse and standard parse. Most data flow components can use only standard parsing; however, the Flat File source and the Data Conversion transformation can use either fast parse or standard parse. For more information, see Parsing Data.
Many transformations compare data values. For example, the Aggregate transformation compares values for the purpose of aggregating values across a set of data rows, the Sort transformation compares values in order to sort them, and the Lookup transformation compares values against values in a separate reference table. To specify how strings should be compared, the transformation includes a set of comparison options such as whether to ignore case sensitivity, how to handle kana types in Japanese text, and whether to ignore white-space characters in the string. For more information, see Comparing String Data.
The expression evaluator also compares data values when it evaluates the expressions that variables, precedence constraints, and transformations use.
Once you have deployed a package to the Integration Services catalog, you can analyze the data flow in the package during execution to check performance or look for other issues. Standard reports are available that allow you to view package status and history, and you can query database views that provide detailed information about package execution. You also can dynamically add and remove data taps during execution to target specific components of your package. For more information, see Analysis of Data Flow.