Kettle experiment conversion case

1. Conversion case

The easiest case to implement is to generate random numbers , And store it to txt file Create a new transformation and save it as test_random（ The suffix is .ktr） Drag and drop plug-ins , In the core object -> Drag input and output respectively “ Generate random number ” and “ Text file output ” Two buttons , And then click “ Generate random number ” And press shift key , Point to with the mouse “ Text file output ”, In order to produce a scissor , Indicates the flow of data . Here's the picture ：

  Edit input stream , namely “ Generate random number ” Button , As shown in the figure

 

2. Conversion case

Then edit the output stream , namely “ Text file output ” Button , As shown in the figure

View results , As shown in the figure

 

 

3. Formal experiment

Through these simple experiments , We know the basic operation of the I / O stream , Next, we enter the formal experiment

1） Change the input and output of the above experiments to Excel. The configuration is as follows ： Excel Input ：

Under file options , The table type is adapted according to the actual situation （xls or xlsx）, After a file or directory , Click on “ Browse ” Choose your own source data file , And then click “ add to ”;

Under worksheet options , Click on “ Get sheet name ...” Add sheet , namely Excel Medium sheet1;

 

  Under field options , Click on “ Get the fields from the header data ...” Get fields automatically , Because of the original Excel The transfer in of middle integer data will become floating point data , So it needs to be changed , As shown in the figure ： 

 

 

 

1） Change the input and output of the above experiments to Excel

The configuration is as follows ： Excel Output ： Just configure the output filename , Others are default .

 

2） The core steps that need to be handled in the next step , That is, switching . First, generate data for 1 Analyze , because kettle The middle group needs to be sorted first , So we need to deal with ：

（1） Cut off the registration date and time to ;

（2） Change age according to certain criteria （ Define your own ）;

（3） Sort by fields to be grouped ;

（4） Group statistics . According to the above ideas , stay “ transformation ” and “ Statistics ” In the core object , Find the corresponding components respectively , Complete basic data flow node configuration , As shown in the figure ： 

 

stay “ Field selection ” In the component , Deal with time . In the metadata options , Need to be right Date Convert to String, Format to yyyy-MM-dd, At the same time, you can rename the field . In addition, you can select fields , modify , remove . As shown in the figure ：

 

 

3） stay “ Range of values ” In the component , Deal with age , The dividing standard defines itself , As shown in the figure ：

 

4） stay “ sort record ” In the component , According to the requirements of generating data , Need to date , Gender , Sort by age , As shown in the figure ：

 

 5） stay “ grouping ” In the component , Group statistics , As shown in the figure ：

6） perform , The result is shown in the figure :