Sunteți pe pagina 1din 2

We have totally 9 test cases that is sequence of testing, 1.data avalibility 2. data load insert 3.

data load update 4. incremental data load 5.data accuracy 6.verify data loss 7. verify column mapping 8. scd type 2 implementation 9. verify naming standards. this r the unit test cases in info we r doing test in manually. this is my concepts if any one others sugg plz forward to me. There are two must test approaches. First one from coding standpoint. You should have this customized as per your org's coding standards. Next one application test case. This should primarily focus on functional aspect of the mapping (what is intended from source to target). In the first place you need to check the source and target definition for correct data type. Then if you have SQL override you need to verify the order of the port. Then business logic. Once the data is loaded you need to verify the record count and quality of data. If you want to check with SQL...you can implement the mapping logic with SQL or Procedure....execute it and compare the data with the data provided by mapping. what we will do in etl testing as a QA , same like writing test cases, executing, all phases of testing . Testing of ELT is slightly different then the Generic testing. Here in ETL testing we need to test the Transformation logic and Business Requirement. For that we need to create the SQL Queries for source level as well as for target. Then we need to compare the result of the both to verify the correctness. Basic checks are: Source and Target record count Source n target data type check Diff Constraints check Data check etc..... I hope this will give u a slight idea..
How can you recognize whether or not the newly added rows in the source are gets insert in the target Based on the mapping logic you have implemented, write simple test cases to validate that the count between the source and target tables is a match. With regards, to new inserted rows in the target, you may have some timestamp field in the target which you can use to match the count of the source and target newly inserted rows. If you want to do a complete data validation between source and target table, you can write MINUS queries based on transformation logic used that will return rows, if there is a data mismatch between source row and target row.

In interview the persion asked test case to check lookup not for complete mapping. If you have some more test cases let me pass. The basic test cases are 1) The input field and the lookup field should have the same datatype and length 2) For the valid input data which is available in the lookup table, it should bring the values .For the values which is not avail in the lookup table ,it will bring the default value or NULL value. 3)If you use SQL override for lookup ,the port order should be same as like in the select query 4) If the input data finds a match for more than one records in the lookup table. The lookup output will be the 1st matched record

S-ar putea să vă placă și