Help with common fields and synthetic keys

[Anonymous]

Hello everyone,

I'm having trouble figuring out how to link my tables together so that I can use one filter to select on all data. I have two common fields, School and Gender in both of my main tables and want to be able to make selections / comparisons across all data.

My current data model looks like this:

I have some bookings data that looks like this:

BookingsKey	Enrol ID	Course Code	Student ID	Gender	Date From	Date To	FlagGL	School
66146-YLE1	66146	YLE1	52003	M	06/07/17	17/07/17	0	L
67541-YBE1J	67541	YBE1J	53282	F	03/07/17	14/07/17	0	B
67561-YBE1	67561	YBE1	53296	M	03/07/17	14/07/17	0	B
69338-YLE1	69338	YLE1	54953	M	06/07/17	17/07/17	0	L
69338-YLE2	69338	YLE2	54953	M	20/07/17	31/07/17	0	L
75823-YBIAE2	75823	YBIAE2	60839	F	03/07/17	28/07/17	0	B
74851-YLE1	74851	YLE1	60090	F	06/07/17	17/07/17	1	L
74306-YBE1	74306	YBE1	59594	M	03/07/17	14/07/17	1	B

Each School has a fixed capacity per gender which is held in a separate table called YL:

YL.Date	WeekYear	YL.School	YL.Category	YL.Gender	YL.Capacity
02/07/2017	27-2017	B	GL Beds	M	2
02/07/2017	27-2017	B	GL Beds	F	2
02/07/2017	27-2017	B	Total Beds	M	150
02/07/2017	27-2017	B	Total Beds	F	159
05/07/2017	27-2017	L	GL Beds	M	1
05/07/2017	27-2017	L	GL Beds	F	4
05/07/2017	27-2017	L	Total Beds	M	99
05/07/2017	27-2017	L	Total Beds	F	154

(The bookings_x_dates table is created by a while loop that generates all the possible dates for each booking using the DateFrom and DateTo fields - )

Ultimately I want to be able to filter on School and Gender to see how many students/GLs we have on any given date/week compared to how many beds are remaining in the school in a table similar to the below:

Obviously one of the problems is that the two bottom rows, no. of GLs and no. of students are not being split by gender because the two tables are not linked properly. If I unqualify the gender and school fields in the YL table though this causes chaos with links and synthetic keys (see below).... How can I get around this problem? Would a link table be the way forward and if so, how would I go about that?

[CarlosAMonroy]

Hi Josephine,

There is a step missing, after doing the interval match you have to join the result with the Bookings table, that way you will avoid the Synthetic table.

Carlos M

[vishsaggi]

Interval Match creates a Synthetic Keys and it is natural. You should be fine having that, and is allowed.

Read here the last two paragraphs.

IntervalMatch

[CarlosAMonroy]

You can try something like:

Bookings:

LOAD BookingsKey,

     [Enrol ID],

     [Course Code],

     [Student ID],

     Gender,

     [Date From],

     [Date To],

     FlagGL,

     School

FROM

[https://community.qlik.com/thread/265110]

(html, codepage is 1252, embedded labels, table is @1);

//Get Max and Min date to create a table with date range

MaxMinDate:

LOAD

  Min(num([Date From],'###0')) as MinDate,

  Max(num([Date To],'###0')) as MaxDate

Resident Bookings;

LET vMinDate = Peek('MinDate',-1,'MaxMinDate');

LET vMaxDate = Peek('MaxDate',-1,'MaxMinDate');

DROP Table MaxMinDate;

//Create a date range

Dates_Tmp:

LOAD

  Date($(vMinDate)+(IterNo()-1),'DD/MM/YYYY') as Date

AutoGenerate(1)

While ($(vMinDate)+(IterNo()-1) <= $(vMaxDate));

//Use interval match to create entry for each date

BookingsTmp2:

IntervalMatch(Date)

LOAD Distinct [Date From], [Date To]  Resident Bookings ; //Important to use Distinct clause other wise the data will be multiplied

DROP Table Dates_Tmp;

//Join Dates to Fact table

Left Join(BookingsTmp2)

LOAD *

Resident Bookings;

DROP Table Bookings;

NoConcatenate

Fact:

LOAD

*

Resident BookingsTmp2

Order by Date;

DROP Table BookingsTmp2;

[Anonymous]

Hmm, I tried the left join but it dramatically increased the number of rows in the bookings table and didn't display the right values in the pivot table. I think I will leave the synthetic key in there for now after reading Henric's post on the topic. Thanks so much for your help everyone, I was really struggling with this one!