On-site solution to project problems-how to check report data

On-site solution to project problems-how to check report data

With the development of business and management, reports are constantly emerging, including finance, operation, statistics, and supervision. Some people laugh at the following: "1/3 of banks are making reports ", it is estimated that 1/3 of companies and developers are switching around users' reports.

When the report was made again, "Let's see if this data is correct. It was correct last month. What happened? Can we still trust this report data? Check the reason and delay the responsibility for submitting the report. "It is estimated that all the users who developed the Report on the site have heard these words. Users are in urgent need of developers. Where can I start to find reports with incorrect data?

In fact, everything is regular, and the report is no exception. To find out the reason for report inequality in a short period of time, you must start with the process, focus on the details, accordingReport process: Choose "source data"> "Data Extraction"> "Data fetch rule"> "Generate Report ".Perform the following steps step by step:

1. Check the extracted data and source data: First, let's take a look at whether there are any omissions or errors in data extraction. Some people tend to ignore this step and think that the data extraction from the previous month is smooth and there is definitely no problem. In fact, this is not necessarily true, some subjects may have no data last month, so the problem is not exposed. Some people will habitually remove the balance of 0 When writing the ETL process. This is not feasible, because some subjects have a balance of 0, but the amount is not 0.

Be sure to check it.Use source data as parentI checked the extracted data one by one. Once I went to the site, the developer told me that I had checked the data. As a result, I found that I missed two copies, the reason is that he used the extracted data as the parent to cancel the core. The data can only be wrong and is missing.

2. Check the number fetch rule:When we confirm that the extracted data is completely correct, we need to check the number fetch rules for the error items. The principle is from total to points, from simple to complex. If there are many tables that are not fair, first check the tables generated by the subjects.

First, check whether the business rule of the number is correct, then check whether the statement matches, whether there are write errors and omissions, and second, check whether there are exceptions in the number rule. For example, year-end carry-over will affect the profit and loss subjects, therefore, the number fetch rules at the end of the year and at the same time are different. This must be noted that the report at the end of the year is the most likely to be uneven, many of which are caused by such reasons. The third is to check whether the boundary value or median value meets the requirements when the fetch rule has a range.

3. Report check:If the preceding two checks are correct, but the report data is still uneven, check the report.

First, check the entire section or the balance sheet. This type of table is based on the subject. The counting rule is simple, and the reason for the imbalance is easy to find out. The entire branch or the balance sheet has been flattened. At least you can confirm that there is no problem with the system's general ledger data. The other is a problem with detailed data.

When you check other reports, you must sort them to check the problem. If the data in the column is not normal, first remove other conditions for query, as shown in the following example:

 

item		A	B	C	d	E	F	I
		Remaining Period						Total
		Next day	2 to 7	October 8 to 30th	31 to 90	91 to 1 year	Over 1 year
	various loans							0.00

1)Total:Without the A-F of any judgment conditions, only determine the subject is the loan details of the data is accurate, if correct, then prove that the total details of the data is no problem, the allocation conditions in the middle of the problem. If it is incorrect, the detailed data extraction is incorrect (this is rare because the detailed data extraction must be consistent with the general ledger data ).

2)Exclusion condition:With the conditions A, B, C, D, and E, only the last condition is left (that is, all the remaining data is placed in the column for more than one year). If yes, the problem is very clear, that is, some detailed data does not meet all the conditions of the A-F.

3)Find data:Add all the conditions for A, B, C, D, E, and F, and add the remaining items (locate the remaining data that does not meet all the conditions ).

4)Analyze the final cause:Identify accounts based on the remaining data to view the cause of the error. The remaining term of the loan is generally offset by the current date and the loan termination date to see if there is a problem with the end date (the date is incorrect or is blank ).

after you have mastered the report checking method, you will find that the effect is remarkable. After a report cycle (monthly, quarterly, or annual, the report data problem is basically solved.