How big Data projects better apply use case Specification Management test Cases

The importance of how a big data project can better manage test cases is self-evident, and one of the most effective methods is a strong and powerful code of execution of use cases;
The use case writing specification is divided into two parts;
The first part: Functional test Case Writing specification
(i) test Case writing specification:
1, the requirement (algorithm) document path;
2, Er-win, data dictionary;

Test Purpose:

Pre-conditions:

Operation Steps:
1.
2.

Expected Result:

(ii) SQL use case writing specification:
1) Each table must use a meaningful alias;
2) When using the table connection, the field to be associated from the table must be placed on the left, the Main table field on the right;
-Requirements, proportions:
The right demonstration:
Select Cv.version_oid, Ce.carid from Gf_mix_city_excle CE
Left JOIN tb_gfda_car_version cv on cv.version_oid = Ce.carid

Error demonstration, the alias of the master, from the table is meaningless, the fields of the associated table are placed on the right:
Select A.version_oid, Dd.carid from Gf_mix_city_excle DD
Left join tb_gfda_car_version A on dd.carid = A.version_oid

2) When a statement needs to have multiple query statements combined with a query, the with structure must be used;
3) each table construct (as) must be followed by a comment, such as: T1 the number of cards in all the cities below;
SELECT *
From (Select Ds.level1_name | | '-' | | Ds.level2_name as Two_segment,
Ds.level2_name | | '-' | | Ds.level3_name as Three_segment
From Dm_segment DS
where Ds.level_flag = 4
ORDER BY Ds.level1_name,
Ds.level2_name,
Ds.level3_name,
Ds.level4_name) T1
where t1.two_segment = ' car-c ';

To modify to the following format:
with T1 as (--T1) The first to second-tier market name is merged as a second-level name; the second to third-tier market name is merged as the third-tier name;
Select Ds.level1_name | | '-' | | Ds.level2_name as Two_segment,
Ds.level2_name | | '-' | | Ds.level3_name as Three_segment
From Dm_segment DS
where Ds.level_flag = 4
ORDER BY Ds.level1_name,
Ds.level2_name,
Ds.level3_name,
Ds.level4_name
)
SELECT * from t1 where t1.two_segment = ' car-c ';

Part II: Data validation use case writing specification
1, Data verification requirements:
-Requirements, background: where the data originated, where, why to clean, what to do, what to achieve the requirements/standards
2, the data source file acquisition path (method);
3, data cleaning process;
--requirements, highlighting the process of data flow, such as after several layers: the original file, DW--DM
4, internal data business process;
--Requirements, for example: the original file provided by the customer to PM--Send by the Business Department audit data Integrity---after approval by the BI group to the original file data directly into the DW, after cleaning, based on business logic to the DM layer)
5. Data cognition Related information:
1) database configuration information;
2) database access account password;
3) data Model (ER-WIN), data dictionary;
4) The data cleaning process involves the table;
6, data update cycle;
--requirements, highlighting the laws of data changes, such as: Deal update frequency: weekly; usually get last week's latest deal price data for every Monday,
7, check the time limit;
--Requirements, for example: In general, the storage is completed in the morning of the week, before work in the afternoon to proofread the completion;

SQL Check case Execution section:
1. Test Purpose:
2. Pre-conditions:
3, calibration method;
--Example: Excel + SQL
4, the operation steps;
5. Expected results:

How big Data projects better apply use case Specification Management test Cases