Design patterns: elements of reusable object-oriented software
Design patterns: elements of reusable object-oriented software
Domain-specific languages: an annotated bibliography
ACM SIGPLAN Notices
Lineage Tracing for General Data Warehouse Transformations
Proceedings of the 27th International Conference on Very Large Data Bases
Practical Lineage Tracing in Data Warehouses
ICDE '00 Proceedings of the 16th International Conference on Data Engineering
Data integration: the teenage years
VLDB '06 Proceedings of the 32nd international conference on Very large data bases
Hi-index | 0.00 |
In order to fulfill its monetary policy function, the Swiss National Bank (SNB) collects statistical data on the economy. The SNB stores results of the regularly held surveys in a specialized database (primary), ordered by surveys and survey forms. After validation the data has to be transferred in another specialized database (secondary) where it can be accessed by economists. The secondary database keeps the data in time series that are hierarchically arranged by statistical taxonomies. The data transfer from the primary to the secondary database feeds 1.5 million time series. Mapping and transformation logic was hard-coded in legacy programs. They were cumbersome to manage and intransparent to the economists in charge. In this paper we describe a novel approach called MapLan, a Java-based data mapping system featuring a domain specific language. The MapLan system not only performs the data transformation and mapping, it also produces complete data lineage information. This paper shows in practice that domain specific languages are an efficient tool to solve two pressing data mapping and transformation problems of statistical databases. One problem is that of mapping the large and heterogeneous schemas of statistical databases in an efficient and manageable way. The other problem is the business need for complete data lineage of the target time series.