Problem Description & Analysis:

The query results of the Postgresql database return 6 columns, with the first column being the grouping column, the second column being the Value column within the group, and columns 3–6 being the Key columns within the group. Each record only has one Key column with a value, while the other Key columns are null. The position of the Key with a value for each record is uncertain.

Task: Now we need to keep the grouping column unchanged, convert the details within the group from rows to columns, convert the values of non-null Key columns to new column names, and convert the values of Value column to new column values.

Solution Highlights:

There are two ways to implement row to column conversion in SQL: one is through functions such as pivot and crosstab, and the other is to use the max group by statement for each new column. However, these two methods require writing new column names, which are the field values of the original table.

When the data changes, SQL also needs to change accordingly, which is very inflexible. If you want a set of code to be applicable to any data, you need to dynamically generate column names using stored procedures or high-level languages, and then execute the previous SQL, which makes the structure much more complex.

The row column conversion function of SPL can be used without writing field values, and a set of code is applicable to any data.

A1: Query the database through JDBC.

A2: Use the pivot function to convert this group of records from row to column. The first parameter is the grouping column, the second parameter is the column name of the Key, and the ifn function takes the first non-null member in the set. The third parameter is the Value column.

---

🚀 esProc SPL FREE Download — Try It Free, Download Now!