The so-called report is the data result presented after various processing and summarization based on the existing data, and the report test is the process of identifying the correctness, completeness, safety and quality of use of the report.
Report test is an important test content, because the user group faced is generally the company's senior management or an important group of users. Problems have a greater impact, so they must be treated carefully and cautiously. This article makes a simple summary based on my previous testing experience and other relevant information.
1. Improve familiarity with the business
Like functional tests and other tests, report tests also need to be familiar with the business, including business processes, business rules, and data storage. The difference is that report tests need to understand the algorithm of each indicator, data sources, and understand the relationship between specific business actions and indicators. relationship between.
2. Data preparation
- Raw data:
1. Understand the table structure\data classification of the original library.
2. Understand the boundary rules displayed in this report, and prepare test data accordingly.
3. Generate data and fix test data by certain means. - Display data:
1. The data covers all categories.
2. The amount of data needs to be large enough.
3. Need to include all boundary values (combined with the query conditions when displaying).
4. The data needs to contain a small amount of illegal data to verify the fault tolerance of the system.
- Data generation method:
1. Stored procedure.
2. Third-party tools (DataFactary, etc.).
3. Generate data through business (not recommended).
4. Related business interface generates data. 3. Data correctness verification
- For customers, the use of reports means that they can quickly and easily find the data they need through the platform of the report system, so the most important content of the test report is to verify the correctness of the data.
- The summary method is as follows:
- To figure out the source of the data, which table and field it comes from;
- Time condition: The statistical interval should be in the card at what time in the business, and consider whether the demand includes the boundary value of the statistical interval;
- To clarify the specific conditions of the measured table and the measured indicators, for example, to count all agency business in this month from 2023-01-01 to 2023-01-31, the specific condition is to limit the business source of the policy to the agency business middle;
- Sql preparation, this process is to summarize the above three processes, and it is also the basis for subsequent data analysis with developers, so improve your ability to write sql. In addition, when the test time is not sufficient, for some simple reports, such as lists and other reports, you can directly select lists of various business types for separate analysis without writing SQL statements by yourself;
- For data checking and analysis, the data queried by sql should be checked with the developer’s. Since some data are large in volume, it is best to use a comparison tool (recommended: BCompare software), and analyze the data that cannot be checked separately , the process of analysis is often the main link to discover problems. In this process, if you really can’t analyze it yourself, you can ask developers to assist;
- Data display format: whether the decimal place, thousandth sign, percent sign, etc. are consistent with those set in the report, whether the unit, exchange rate, etc. are converted, and whether some codes are converted into text;
- Consistency between detail and total, whether the sum of the detail values of each part is consistent with the total, etc.;
- To cover all query statistics methods, under the condition of sufficient time, it is necessary to set various condition combinations through equivalence class division and permutation and combination according to the conditions (filter items, dimensions). the custom shall prevail;
4. Display of report format
- After data verification, the focus is whether the display format of the output report meets the customer's needs.
- There are two main types of report formats
- The reports stipulated in the insurance industry standards use a fixed format, such as: some reports reported by the China Insurance Regulatory Commission,
- Reports are customized according to the needs of enterprises or users, so these two types of reports need to be tested from the following aspects:
- Whether the overall display format of the report conforms to the form provided by the customer;
- Whether the title or table name of the report is correct;
- Whether the time period of the report page is the time period selected by the user;
- When the output content is too much, whether the pagination method is correct, whether it has the same style as the previous page (such as the header) when turning the page, and whether the output of the second page is correct;
- Whether the data that needs to be reminded (some abnormal data) is highlighted, and whether some indicators are calculated in a particularly complicated way or if there are several indicators that are easy to confuse, whether there are comments on the page;
5. Comparability between reports
- After the vertical test is completed, we need to connect the tested reports horizontally, because some reports have different names, but some indicators are the same, so we need to compare the two reports to see if they are the same Whether the statistical results are the same within the time period.
- In addition, different indicators in different reports are also related, such as: the list of premiums receivable in business and the balance of premiums receivable in finance. When the two statistical standards are consistent, the total of premiums receivable in the list is It is equal to the balance of financial receivables, and there are also premium income, actual premiums received, and premiums receivable in the same statistical interval. Total premium income = actual premiums received + premiums receivable (not actually received). To sort out the levels and sequences between them, it is necessary to strengthen the understanding of business and the accumulation of knowledge!
6. Other
- Report output and printing:
After the report is generated in the system, it does not end. Generally, the report needs to be printed for customer use, such as meeting or submitting for approval. Therefore, the printing function of the report is also very important. Before printing, Users generally need to export the report for further analysis or for comparison with other reports. Therefore, the export function of the report must also be verified. Generally, the main formats that can be exported are Excel and pdf formats, and then it is necessary to verify whether the exported content is correct. The generated reports are consistent. - Report performance: Try to require developers to use the optimal query statement to avoid customers waiting too long during use
- The authority of the report; for a system with authority control, the report should of course be consistent with the authority of the user. It is necessary to verify the authority control from two aspects. Definition of report conditions: in the condition selection area, some drop-down boxes should not be able to Display data outside the scope of the user's permissions.