Don't join - use Applymap instead

[hic]

My latest blog post was on joins in QlikView (see To Join or not to Join). In it I claimed that you should avoid making joins in the QlikView script, if possible. This blog post is about a function that can help you avoid joins. It is about the function Applymap().

[hic]

It is, in fact, one of my favorite functions and I do not seem to be alone in this choice. Several of the QlikView developers who regularly write on QlikCommunity seem to share this preference and have written about this function, e.g., QlikView Notes, Quick Intelligence and QVApps.

So, what does the function do? Well, basically it is just a lookup function – it takes one value as input, checks if this value exists in a mapping table and returns the corresponding value from the second column in the mapping table. Think of it as a translation function where the translation is fetched from a pre-defined, two-column table. A simple translation function, but you can still do a lot with it...

What has this got to do with joins? Everything. A very common situation in data modeling is that you need to get a single field value (per record) from a different table than the current one. Then the standard way to do this is to use a join. However, in QlikView you can – and should – use Applymap() instead. The only time that you cannot use Applymap() instead of a join is if you want to fetch more than one single corresponding value (record) from the second table.

Let's look at an example of how to use Applymap(): In an ERP system, the customer table is the table where all information about the customer is stored, e.g. the name and country of the customer. Further, it could be that you need the customer’s country when making a calculation in the order table. This is a field that you don't have in the order table. In SQL you would join the two tables to get country into the order table.

In the QlikView script you would instead first define the mapping table that maps a customer ID to a customer country as follows:

     MapCustomerIDtoCountry:
     Mapping Load CustomerID, Country From Customers ;

The second step is to use this information when reading the order table, such as:

     Orders:
     Load *,

          ApplyMap('MapCustomerIDtoCountry', CustomerID, null()) as Country
          From Orders ;

The "null()" as third parameter of the function defines what it should return when the customer ID isn't found in the mapping table.

And with this, you have joined the field Country with the order table – without using a join. And you have done it faster and with less chance of errors. Bottom line: Whenever you know that you just want to fetch one single value per record – don’t join. Use Applymap() instead.

HIC

See more in the Technical Brief on Joins and Lookups.

[hic]

I might use the Join prefix for small tables, but not for big since it uses a lot of memory. So, the bigger table, the more reason to avoid the Join prefix.

In your case, I would ask myself whether there is a risk with joining (record duplication), and if not, if I can do it on the database (using a JOIN inside the SELECT). Because that could be an option.

But if you cannot do it on the database, I would stick to 5 mapping tables.

Finally, we should always ask ourselves: What am I trying to optimize? Here the answer is probably script execution time. Next question is whether script execution time is a problem. If it isn't and I have a functioning solution, well then  ... I personally often find myself optimizing things that don't need optimization. Because it is a fun challenge...

HIC

[#4]

gr8 gr8 gr8!!!!

[#5]

Henric,

THIS

MADE

MY

DAY (or week)

I was crashing my head to have an intricate join yield unique values, where duplicates where coming up and expected due to the nature of the tables...

This solution is absolutely simple, lean, fast, harmless and PERFECT!

just one question, of which I'm 99% sure of the answer.

When the Mapping table has

CustomerID     Country

2                    Germany

2                    France

The function is picking up Germany I suppose, i.e. the first correspondence found for the mapped key, right?

Again thank you!!

[hic]

Glad I could help you.

And, yes, you are right about duplicates in the mapping table. For the sake of simplicity, it picks up the first occurance.

HIC

Read more on my other blog posts.

[kalyandg]

Dear Henric,

Thank you for posting such a useful info, am doing project, in which i have linked five tables with one table.

for example, the script as follows in my project

Master:

Item,

Product,

ProductGroup,

ItemLocation

Size

ProductName,

Level1:

Item,

L1Code

Level2:

Product,

L2Code

Level3:

Product Group,

L3Code

Level4:

ItemLocation,

L4Code

Level5:

Size,

L5Code

Level6:

ProductName,

L6Code

these all links each with corresponding fields in master table

but for optimizing application, i have to use applymap and mapping load

please help me to solve this, and provide me a script code for each level and how to map as in my script.

[hic]

If you want to include the L-codes into the master table, then it is straightforward. First create a mapping table:

          'L1Map':

          Mapping Load Item, L1Code From Level1;

and then use this information in the Master table load statement.

          ApplyMap('L1Map', Item) as L1Code

The same should be done for all 6 levels, i.e six different ApplyMap() calls, each creating a new field.

If you have more questions on your specific problem, I think it would be better to enter it as a question on our discusion forum.

HIC

[kalyandg]

Dear Henric,

Thanks for your reply i have one doubt wheather we have to  load Master table first or Level table and also we have more than two fields in level 2 table what we have to do? 

[hic]

You have to do the mapping loads of all levels before you load the master table with Applymap.

The mapping load can only have two fields, but the table you load from can of course have more. Just make a load statement with the two correct fields and it will work.

HIC

[jesper_bagge]

Hi Henric,

Whats your take on using 5 Lookup() against the one table instead of creating 5 mapping loads and 5 ApplyMaps()?

Is there performance to be won in using ApplyMap() instead of Lookup()?

Regards

Jesper

[hic]

The performance of applymap is much better, so I would use 5 mapping tables and 5 applymap calls.

HIC