Title: Testing Framework for Query
1(No Transcript)
2Effectively Validate Query/Report Strategy and
Tool
Session id 36993
- Steven Luo
- Sr. System Analyst
- Barnes Noble
3Agenda
- PART I Introduction
- Why, when, and how to validate
- PART II Strategy
- Basic strategy
- Advanced strategy
- PART III Tool
- Test script
- Engine - sqlUnit
- PART IV Conclusion
4Why to Validate?
- Assure Data Accuracy
- Managers need accurate data to make strategic
decisions - A companys sales team needs accurate data to
launch market campaign - Assure ETL (Extraction, Transformation, and
Loading) Process Correctness - ETL tools
- SQL scripts
5When to Validate?
- Any of the following types of testing needs
validation - Unit testing
- Integration testing
- System testing
- Acceptance testing
- Maintenance regression testing
6How to Validate?
7Agenda
- PART I Introduction
- Why, when, and how to validate?
- ?PART II Strategy
- Basic strategy
- Advanced strategy
- PART III Tool
- Test script
- Engine - sqlUnit
- PART IV Conclusion
8Validating Report/Query
- A Query or Report is actually a result set, so
validating needs to answer two questions - Are you getting the result set right?
- right data in each cell
- Are you getting the right result set?
- exact number of records
-
9A Typical Process of Generating Reports
ETL steps
Stored procedure
report
staging tables
materialized view
flat files
T0
T1
Tn
...
10Basic Strategy
- Assert that each (cell) in the result set matches
the data in the source table - Number
- String
- Date
- Result Set, etc.
- Sampling should be used if a result set is big
- Assert the right number of records in the result
set - Check duplicated
11Advanced Strategy
- Try to uncover invalid data item
- Boundary Validation
- Count, sum, max, min, x not in table,
between..and on whole resultset or certain
partitions. - Special Value (constraints) Validation
- isXXX() and notXXX()
- e.g. IsNull, notNull, notNegative, notZero
- Business Rule Validation
- dept1.salegt dept2.sale
12Agenda
- PART I Introduction
- Why, when, and how to validate ?
- PART II Strategy
- Basic Strategy
- Advanced Strategy
- ?PART III Tool
- Test Script
- Engine - sqlUnit
- PART IV Conclusion
13Why Use a Validation Tool?
- Automate the validating process
- Reduces the cost, time and effort
- Reuse the procedures
- Write once, run many times on QA box and/or
production box. - Re-factor SQL
- Share by group (save to PVCS)
14Tool
- Open Source
- jUnit, etc.
- Steven Feuersteins utPL/SQL
- My tool -- sqlUnit
15sqlUnit Overview
- The framework consists of two major parts
- test scripts
- test engine
- implemented in java stored procedure with PL/SQL
interface - Implement 2 types of strategy
- basic strategy
- advanced strategy
- Record the validating results
- Monitor long-running validation process
16sqlUnit Overview (2)
- From the Engine Perspective
- run all your test cases defined in your PL/SQL
package - From the User Perspective
- write all test scripts
- start the Engine
17Architecture
Engine
Assert
Test scripts
util
Database
18Test Script (1)
- Write Test script in PL/SQL
- Define test package
- Call APIs
- Define test procedures
- Test procedure MUST begin with test
- Setup()
- Teardown() clean up
19Test Script (2)
- Use your business knowledge to get expected data
and actual data - Get raw or original data from source table
- such as POS, Daily Sales, etc.
- Get data from a report
- call API fetchCursorData(), or
- using cursor directly, e.g.
cc my_test_pkg.get_ref_cursor('SCOTT') loop
fetch cc into value1,value2,..., valuek
exit when ccnotfound if(...) then
sqlunit.assert(desc, value1, 100) end if
end loop
20APIs for Basic Strategy
procedure runTestCases(testPackageName
varchar2) procedure assert(description varchar2,
num1 number, num2 number) procedure
assert(description varchar2, str1 varchar2, str2
varchar2) procedure assert(description varchar2,
a1 STRING_ARRAY, a2 STRING_ARRAY) procedure
assert(description STRING_ARRAY, a1 STRING_ARRAY,
a2 STRING_ARRAY) procedure assertQuery(description
varchar2, query1 varchar2, query2 varchar2)
procedure fetchRefCursorInto(pname in varchar2,
parameters in
STRING_ARRAY,
fetchfields in out STRING_ARRAY,
uniqField in varchar2,
uniqValue in varchar2 )
function getCountForQuery(sqlstr varchar2)
return number function getCountForProcedure(sqlstr
varchar2) return number
21Test Script Template
22Apply Basic Strategy
- A tool should pick up
- m column(s) in a row of a result set
- m between 1 and number of columns
- m column(s) in n rows of a result set
- m between 1 and number of columns
- n between 2 and number of rows
- Cells to be validated
- cells between 1 and mn
23(No Transcript)
24Apply Advanced Strategy
- Advanced Strategy
- Boundary Object
- Special value (Constraint)Object
- Business Rule Object
25Test Script for Advance Strategy
procedure test_adv_1 is obj
sqlunitBoundary sqlunitBoundary(NULL,NULL,NULL,
NULL) begin obj.setTestingQuery('my
_test_pkg.get_ref_cursor(''SCOTT'')')
obj.setCountCriteria(2)
obj.setQueryCriteria('c1 is not null')
obj.setQueryType(1) //1 store procedure. 0
sql query obj.checkBoundary
exception when others then
dbms_output.put_line('exception! ')
end test_adv_1
26Example for Validating Two Queries
27How Engine Works
- Users start the Engine by calling
- exec sqlunit.runtestcases('VALIDATEPACKAG
E') - Engine calls back test scripts by calling the
following - set_up test_1 tear_down
- set_up test_2 tear_down
- ...
- set_up test_n tear_down
- Test scripts call Framework APIs .
- sqlunit.assert(...)
- sqlunit.assert(...)
- sqlunit.assert(...)
28Sequence Diagram
testPackage
sqlUnit/engine
DB
Runtestcases()
setup
test_1
assert
recordResult
tear_down
setup
test_2
recordResult
assert
tear_down
View results
29Record the Test Result
- Use package name as testing result table name.
30(No Transcript)
31Monitoring Validation Process
- At Engine level, by instrumentation
DBMS_APPLICATION_INFO into Engine, you can
monitor the progress - At test script, you can instrument
DBMS_APPLICATION_INFO to test scripts too. - Get the progress information from
vsession_longop in other session
32Test Scripts Guideline
- Be simple
- Dont use the same sql that generates the report
- Use Business/domain knowledge
33Conclusion
- Data accuracy is very important
- Use automated validating tool whenever possible
34Limitation of validation
- Program testing can be used to show the
presence of bugs, but never to show their
absence - --E. W.
Dijkstra
35References
- Asimkumar Munshi, Testing a Data Warehouse
Application white paper http//www.wipro.com/insig
hts/testingaDWApplication.htm - B. Hailpern and P. Santhanam Software
debugging, testing, and verification IBM System
Journal Vol. 41, No. 1, 2002 - Thomas Kyte, Expert One on One Oracle, Wrox,
2001
36A
37(No Transcript)