QV15 Hierarchy In QLIKVIEW

QLIKVIEW wow, a great tool again. Solve the big problem of hanling the hierarchical data.
We have two function at all, which can be used at script level
 - hierarchy(upto 8 variables can pass)
 - hierarchybelongsto( 6 variables can pass.)

both function can use in front of LOAD statement or SELECT.

hierarchy:
it convert the node adjacent table to , in that way each level of child record will written in separate field.

Lets do some handsome practice on table - scott.emp !!!!!!!!

create connection
load the table
below the script you direct copy paste to your script editor after adding connection, i have three set of queries we will see the behaviour of each set of queries one by one:

//HIERARChy(EMPNO,MGR,ENAME)
//LOAD EMPNO,
// ENAME,
// JOB,
// MGR,
// SAL;
//SQL SELECT *
//FROM SCOTT.EMP;

 :::
HIERARChy(EMPNO,MGR,ENAME)
it has three fields:
1. field1 i.e. empno - that has child records only
2. field2 i.e. mgr - reference to parent records
3. field3 i.e. ename - any column name can be used, depend on us what type of data we want to show in hierarchical manner.
 here i want to see the employee name in hierarchical order.


HIERARChy(EMPNO,MGR,ENAME,asdf,ENAME,[hierarchy_goeg], ';', 'HIERARCHY DEPTH')
LOAD EMPNO,
ENAME,
JOB,
MGR,
SAL;
SQL SELECT *
FROM SCOTT.EMP;

:::
HIERARChy(EMPNO,MGR,ENAME,asdf,ENAME,[hierarchy_goeg], ';', 'HIERARCHY DEPTH')it has three fields:
1. field1 i.e. empno - that has child records only
2. field2 i.e. mgr - reference to parent records
3. field3 i.e. ename - any column name can be used, depend on us what type of data we want to show in hierarchical manner.
 here i want to see the employee name in hierarchical order.
4. asdf - i use any dummy name .. either we can use the parent record description related field
5. ENAME - PathSource: The path in QlikView is a string containing one field per ancestor down to the node.
6. PathName: the name of the field that will contain the path.
7. Delimiter: the letter to separate the different fields
8. HIERARCHY DEPTH -the name of the field that will contain the depth of the node.





hierarchybelongsto:
The hierarchybelongsto prefix is used to transform a hierarchy table adjacent node table. Adjacent node table is the table where each record corresponds to a node and has a field that contains a reference to the parent node. (means table will store data in parent child relation help of two fields, parent record repeated over row untill all child has been written corespond to parent. see the below image first two field behaviour, it is a adjacent node table)
:::
HierarchyBelongsTo(EMPNO,MGR,ENAME, 'ANCESTORS_KEY','ANCESTORS_NAME', 'Depth')
it has three fields:
1. field1 i.e. empno - that has child records only
2. field2 i.e. mgr - reference to parent records
3. field3 i.e. ename - any column name can be used, depend on us what type of data we want to show in hierarchical manner.
 here i want to see the employee name in hierarchical order.
4. write parent record in field
5. write child record in field
6. Depth - depth number of record (0 for child node, summing up with 1 for roots)


geo2:
HierarchyBelongsTo(EMPNO,MGR,ENAME, 'ANCESTORS_KEY','ANCESTORS_NAME', 'Depth')
LOAD EMPNO,
ENAME,
JOB,
MGR,
SAL ;
SQL SELECT *
FROM SCOTT.EMP;

 

QV14 Peek() used in for loop

Peek() function can also most usable in for loop in qlikview.
Lets have a look on a little example, I have google-ed this example and perform on my local machine.

In the example, we have a table "FileListTable:" with two fields Date1 and FilNme. FilNme i.e. Filename all are exists on the loacal harddisk, we have to load each file and we can perform any type of string operation on loaded file.
In for loop:
1. NoOfRows return total number of rows in table
2. for loop work from bottom to up. As the logic says "vFileNo-1", Total no. of rows-1
 i.e.  11-1 = 10 return -> Airline Operations_ch8.qvw.2014_07_28_12_18_30.log
   10-1 = 9   Airline Operations_ch8.qvw.2014_07_28_11_34_41.log
   9-1 = 8    Airline Operations_ch8.qvw.2014_07_25_16_51_02.log
   ....
   2-1 = 1    Airline Operations_ch8.qvw.2014_07_25_15_38_06.log
   1-1 = 0    Airline Operations_ch8.qvw.2014_07_25_15_36_35.log

3. In loop we are creating New file name which will same the log file in the form of text file.
4. Store all log file in form of "*.txt"

FileListTable:
Load * inline [
Date1, FilNme
18-07-2014, Airline Operations_ch8.qvw.2014_07_25_15_36_35.log
25-07-2014, Airline Operations_ch8.qvw.2014_07_25_15_38_06.log,
26-07-2014, Airline Operations_ch8.qvw.2014_07_25_15_39_08.log,
27-07-2014, Airline Operations_ch8.qvw.2014_07_25_15_50_00.log,
19-07-2014, Airline Operations_ch8.qvw.2014_07_25_15_50_37.log,
20-07-2014, Airline Operations_ch8.qvw.2014_07_25_15_53_15.log,
21-07-2014, Airline Operations_ch8.qvw.2014_07_25_15_57_28.log,
22-07-2014, Airline Operations_ch8.qvw.2014_07_25_16_48_30.log,
23-07-2014, Airline Operations_ch8.qvw.2014_07_25_16_51_02.log,
24-07-2014, Airline Operations_ch8.qvw.2014_07_28_11_34_41.log,
28-07-2014, Airline Operations_ch8.qvw.2014_07_28_12_18_30.log
] ; 


For vFileNo = 1 to NoOfRows('FileListTable')
      Let vFilNme = Peek('FilNme',vFileNo-1,'FileListTable');
       xyz: Load *,'$(vFilNme)' as FilNme
       From [$(vFilNme)];      
       let vfilenme =replace(vFilNme,'.log','.txt') ;
       store * from xyz into $(vfilenme);
Next vFileNo
 

QV13 Peek() and previous()

the Peek() function allows the user to look into a field that was not previously loaded into the script
whereas the
Previous() function can only look into a previously loaded field.

Both are Inter Record Functions.These functions are used when a value from previously loaded records of data is needed for the evaluation of the current record.  IT loads the value from currently evaluating column. which is being in process and look into the previous value for that column.

peek(fieldname [ , row [ , tablename ] ] )
Returns the contents of the fieldname in the record specified by row in the internal table tablename. Data are fetched from the associative QlikView database.
Fieldname must be given as a string (single quote).

Row must be an integer.

• 0 denotes the first record,
• 1 the second and so on.
• Negative numbers indicate order from the end of the table. -1 denotes the last record read.
• If no row is stated, -1 is assumed.
   

If no tablename is stated, the current table is assumed.

previous(expression )
Returns the value of expression using data from the previous input record. In the first record of an internal table the function will return NULL because no previous record exists for first value.

The previous function may be nested.
Data are fetched directly from the input source, making it possible to refer also to fields which have not been loaded into QlikView, i.e. even if they have not been stored in its associative database.

Some similarities and differenceThe Similarities
 - Both allow you to look back at previously loaded rows in a table.
 - Both can be manipulated to look at not only the last row loaded but also previously loaded rows.
The Differences - Previous() operates on the Input to the Load statement, whereas Peek() operates on the Output of the Load statement. (Same as the difference between RecNo() and RowNo().) This means that the two functions will behave differently if you have a Where-clause.
 - The Peek() function can easily reference any previously loaded row in the table using the row number in the function  e.g. Peek(‘Employee Count’, 0)  loads the first row. Using the minus sign references from the last row up. e.g. Peek(‘Employee Count’, -1)  loads the last row. If no row is specified, the last row (-1) is assumed.  The Previous() function needs to be nested in order to reference any rows other than the previous row e.g. Previous(Previous(Hires))  looks at the second to last row loaded before the current row.

So, when is it best to use each function?    The previous() and peek() functions could be used when a user needs to show the current value versus the previous value of a field that was loaded from the original file. 
    The peek() function would be better suited when the user is targeting either a field that has not been previously loaded into the table or if the user needs to target a specific row.

source http://community.qlik.com/blogs/qlikviewdesignblog/2013/04/08/peek-vs-previous-when-to-use-each

//example1

DR:
LOAD F1, F2, Peek(F1) as PeekVal, Peek('F1',1,'DR') as PeekVal1, Previous(F2) as PrevVal ;// Where F2 >= 200;
LOAD * INLINE [
F1, F2
A, 100
B, 200
C, 150
D, 320
E, 222
F, 903
G, 666
];

//example2

data of file "9_Peek vs Previous.xlsx"
Date Hired Terminated
1/1/2011 6 0
2/1/2011 4 2
3/1/2011 6 1
4/1/2011 5 2
5/1/2011 3 2
6/1/2011 4 1
7/1/2011 6 2
8/1/2011 4 1
9/1/2011 4 0
10/1/2011 1 0
11/1/2011 1 0
12/1/2011 3 2
1/1/2012 1 1
2/1/2012 1 0
3/1/2012 3 1
4/1/2012 3 4
5/1/2012 2 1
6/1/2012 1 0
7/1/2012 1 0
8/1/2012 3 0
9/1/2012 4 3
10/1/2012 0 3
11/1/2012 2 1
12/1/2012 0 0
1/1/2013 4 2
2/1/2013 2 2
3/1/2013 3 0

[Employees Init]:


LOAD
rowno() as Row,
Date(Date) as Date,
Hired,
Terminated,
If(rowno()=1, Hired-Terminated, peek([Employee Count], -1)+(Hired-Terminated)) as [Employee Count]
From
[..\datafiles\9_Peek vs Previous.xlsx]
(ooxml, embedded labels, table is Sheet2);

[Employee Count]:
LOAD
Row,
Date,
Hired,
Terminated,
[Employee Count],
If(rowno()=1,0,[Employee Count]-Previous([Employee Count])) as [Employee Var]
Resident [Employees Init] Order By Row asc;

Drop Table [Employees Init];

image explanation
1. hierd - no of employee hiered
2. Terminated - no. of emp terminated
3. [Employee count] - accumalation(no of working employee - terminated employees)
we used peek function to get the desire result
4. [Employee Var] - employee variation from field [Employee count] using previou().

//now if you observe that we use previous() in different load query, because previous work on pre-loaded field.
//If we try to use previous() in load table "[Employees Init]" then it will not work, because "[Employee Count]" is being calculated in "[Employee Count]" load statement.

QV12 Above() Below() function in qlikview

above and below lie in the Chart Inter Record Functions category. These functions can only be used in chart expressions.

SYNTAX is:
above([ total ] expression [ , offset [,n ]])

It return the value from expression, which is previously seen in the execution.
In other words it return the value the row appear above the current row in table chart.

where we can use above function
- need to create pivot table [accumalation of pivot table]
- need to get the difference and addition of sales (diff. according to month year any perspective) i.e. accumalative sum
Below is the example with some sample data and we will see the behaviour of above() function in straigth table as well as in pivot table.

ex:
LOAD * INLINE [
sno, no
1,20
2,578
3,344
4,54
5,34
6,34
7,34
8,34
];

In chart properties add the below expressions:
- sum(no)
- above(no) ~~ above(no,1)
- below(no,1)
- sum(aggr(rangesum(above(total no)),sno))

Reload the document and see the behaviour of table chart.

image



picturization:
1. sum(no) - will give total sumup of expression
2. above(no) ~~ above(no,1)
as you can see i have written above(no) ,above(no,1) two different entities and but working is same for both. If we skip the offset by default qlikview assume 1.
And see the image above the behaviour is as it start write value from 2 row, by skipping 1, because offset number is one. If we replace it with two it will start writing from 3rd row.

- and you can see the first row having null value, because it there is no value above 20.
3. below(no,1) - It start writing in the column the value below in current expression.

Now we can see the image that the column above(no) and below(no) not summing up the column.
4. to get the sum for the column we need to do "rangesum" with "aggr()" function.



Above() behaviour with pivot table:
script the below inline table

abve:
LOAD * INLINE [
Customer, FK, Sales
A, 1, 50
A, 2, 45
A, 3, 25
A, 4, 75
B, 1, 35
B, 2, 75
B, 3, 12
B, 4, 14
C, 1, 12
C, 2, 50
C, 3, 45
C, 4, 65
D, 5, 45
D, 6, 76
D, 7, 12
E, 1, 12
E, 2, 15
F, 1, 25
F, 2, 56
F, 3, 87
F, 4, 98
F, 5, 54
G, 1, 12
G, 2, 34
G, 3, 15
H, 1, 15
H, 2, 65
H, 3, 15
H, 4, 12
];


LOAD * INLINE [
TestVal
1
2
5
6
7
];

Dimension: - Customers, FK
Expressions:
- Sum (Sales)
pivot_column - =above( column(1))
total_sum - =Above(TOTAL Column(1)) + If(Above(TOTAL Column(3)) > 0, Above(TOTAL Column(3)), 0)

sum customer wise - rangesum(ABOVE(column(1), 1, ROWNO()))

reload and create a pivot table chart.

QV11 Only() function in qlikview

Only function work return the only one value when the value in parameter in common with respect to dimension else it will return null. It always return single value. 

pricem:
Load * inline [
Month, Price
1, 345
1, 342
1, 34
1, 345
1, 67
2, 76
2, 87
2, 67
2, 33
3, 321
3, 123
3, 324
3, 274
4, 33
4, 49
4, 192
5, 923
5, 347
5, 43
5, 84
5, 52
5, 55
6, 10
6, 12
6, 554
6, 546
7, 22
7, 10
7, 20
8, 30
8, 40
9, 50
10, 6
10, 60
10, 70
11, 80
12, 90
12, 90
];


load Month, Only(Price) Resident pricem group by Month;


Using above query it will return value for 9,11,12 month only

This is where the Only() function affects the calculation without the user knowing it; if there is no explicit aggregation function in the expression, QlikView uses the Only() function implicitly. Hence, in some cases,

- “Only(Date)” is used as sort expression
- “=Only(Product)='Shoe' ” is used as search criterion.

case will be updated soon