Create a new transformation and add a Table Input step that can connect to the information_schema database.

Create a query that selects the TABLE_NAME and COLUMN_NAME columns from the COLUMNS table, making sure to filter only on the books TABLE_SCHEMA.

Add a constant value using the Add constants step found under the Transform category. The value should be named grouper with type Integer and value of 1. Create a hop from the Table Input step to the Add constants step:

Add a Denormaliser step found under the Transform category. The Key field should be the grouper column created in the last step. The Group field should be TABLE_NAME. Fill in the Target fields: information like the following:

Preview the Denormaliser step. For each table in the books database, you should see a record with a comma-separated list of column names.

Now finish this transformation by adding a Copy rows to result step and create a hop from the Row denormaliser step to the Copy rows to result step.

Since we will be building a SQL query from these columns, the simplest way will be to use them as variables. Variables can not be used in the same transformation as they are set, plus we will have multiple sets of variables, so we need to create a sub job and a parent job. Sub jobs are jobs within other jobs. Let's continue building the transformations needed and then we will build the two jobs required to run this process.

Create a second transformation. Add a Table Input step that will use the variables we will be creating from the data in the first transformation. Be sure to select the Replace variables in script? checkbox. The query should look like the following:

SELECT ${column_list_par}
FROM ${table_name_par}

Add a Text file output step. For the Filename field, point to a location where the database table extracts can be dumped to. The Filename can also use parameters. Use the table_name_par as the file's name. The Text file output step will store whatever data is in the stream without declaring any Fields, so leave the Fields tab empty.

Create one last transformation that will use the Get rows from result step and load the variables column_list_par and table_name_par with the Set Variables step. The Set Variables step should be filled in like the following:

Create a job. This will be the sub job that will take each record and execute a query. Bring over a START step, two Transformation steps, and a Success step. The first Transformation step should point to the transformation that sets the parameters used in the query. The second Transformation step should point to the transformation that uses the parameters and extracts the query output into a text file. This job should look similar to the following:

Create another job. This will be the parent job to the whole process. Bring over a START step, a Transformation step, a Job step, and a Success step. The Transformation step should point to the transformation that is extracting data from the information_schema database. The Job step will be pointing to the job we created previously. Make sure that the Execute for every input row? checkbox is checked under the Advanced tab for the job. Your final job should look similar to the following:

Now execute the parent job. There should be a number of text files with an output equal to the number of tables in the books database.