Head Tail and Sample

Head

Head is a debug stage that picks top ‘n’  records from all the nodes  (or)Specific nodes.

   

Head  properties

Rows

All rows  (After skip)  = false

Number of rows (preparation) = 2

• Available properties period preparation skip the preparation
• Partitions

All Partitions = True

Click on output Do the mapping Ok  

Do you want to master DataStage? Then enrol in "DataStage Training" This course will help you to master DataStage

In the output:-

The top 2 records will be displayed

[In Sequential file, if we load multiple files, then set

S.F properties àoptions  Keep file partitions = True]

Tail

Tail is a debug stage that picks bottom ‘n’ records from all nodes or specific nodes

• Tail
•  properties
•  Rows

Number of rows (preparation) = 2   Partitions

All Partitions = false

Click on output mapping  

Display last 2 records

Ex:- Suppose that we have 60 records and we need to display the 30th record to the 40th record.

Sol:  In Head  properties Set No. of row  (per partition) = 40

In TailPropertiesSet No. of row  (per partition) = 10

   

Sample

The sample is also a debug stage, which operates in ‘2’ Modes  

• Period  1 input and 1 output
• Percentage  1 input and 1 output

Period is interval of records if period = 2 , displays 2,4, 6,8,10……….

Sample

Properties

Options

Per cent =?

Out click number =?

Sample mode – period /per cent

If period

  Options

Period (per partition)  =2  output à Mapping  ok

If per cent

Since it has multiple outputs it follows the link to order.

Options

Percent = 50

Output link Number =

Percent = 25

Output link number =

Sample Mode = percent  Click on link ordering (To know the link label) 

0       t2

1  ↔  t1

2   ↔  t3

Now assign these values accordingly in output – link number

     ↓

Click on output

     ↓

Perform the mapping  fx t1, t2, t3

     ↓

Ok  

Peek:- (Multipurpose)

Peek is a debug stage that can be used for

• Send the records to the logs
• Make the Source to Multiple copies (can be used as “copy” stage)
• Used as STUB stage (place holder)

Log  Detailed info abt execute

     

S.F  load peek  we can view data here    

Properties

   ↓

No. of Records = 5  (per partition) (All the 5

   ↓

Records will be sent to the job log)

   ↓

Peek Records output Mode = Job log

   ↓

ok   After Compilation and Run, Go to “Tool”

• Tool

      ↓

Run Director

      ↓

Select the job

      ↓

Click on job log Symbol

• In that, we can see 3 types of colours
• Green – ok
• Red – Error
• Yellow – Warning

To clear the Warnings

Right click on Warning (or) double click  Warning   Message will be displayed Identify the Warning and try to Eliminate them

Note:-   “Missing” record delimiter “/r/n”  (one type of error) (it means Mismatch)

In which Cases, we have to put a “Reject link” and capture the mismatched data

Message Handing

Right-click on  Yellow Warning Add to Message Handler  Actions (Suppress / Demote) Add Ruler ok Warning get converted into Information

Filtering:-

Job Right-click filter (put the options)

click on job name  we can see 5 records  

Peek properties 

Rows  all the rows  (after skip)   = false                                         properties

Number of records  (per partition) = columns                              period (per partition)

Peek all input  columns = True partitions                                      Skip  (per partition)

All partition = true

Options

Peek Records output Mode == Joblog                                        properties

(or) output                                                                                    Deli meter string

Show column Name = True

Ex:

Suppose we want to see the data at each stage then attach a peek

Stage to each Transformation :   Oracle --- Sort  Transformation1 Transformation 2 Transformation 3 Data set

                         ↓                      ↓                    ↓

Peek 1                                               peek 2                                        peek 3  

• To select only particular columns:-

Peek can send data to logs, peek can also send required columns to the logs.

For Ex:-  If we have a table that will field and 9 records, and wanted to display only 2 fields with 9 records.

Peek  properties  No. of Records = 4 (per partition) columns Input column to peek = c id, Input column to peek = State, Peek all input columns = false  ok After compile and Run Tools  Run Director Select job  log Event Detail window appears

Output :-  peek 1, 0: cid 1 , state AP

If we want to display only records without displaying column names

Columns Show column Name = false  

• Make the source into Multiple copies

       

 Peek  properties  No. of Records = 5   peek Records output Mode = job log Go to “output” tab  Map the data (l1, l2, l3)  ok

Compile and run the job Here peek Stage ads as copy stage and copies, The data as it is to the 3 data sets.

If we set  peek Record output Mode = output, then

     

Peek  properties  No. of Records = 5  peek Record output Mode = output Output  Map the data  ok  

When we set peek Record output Mode = output, then the peak output

That is 5 records will be displayed in one of the .ds files, and in the other .ds file, All records will be displayed.  

In Data set 3  view data we can see 5 records ( that is peek data)

Suppose the client needs only rejected, but not output, thenHere we use peek as the output stage

   

Here peek is used as STUB Stage, a place holder to store the data.