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.

[Anonymous]

I use ApplyMap when is relevant. It's more confortable and easy to find in the script. Obviously when the source tables are not big. If its limits were eliminated maybe JOIN will be almost forgotten. Thank to ApplyMap creator. CB.

[m2r4miller]

I think 5 ApplyMaps is the way to go, but don't hit the database 5 times - instead load the key field and the 5 other fields from the DB into a temp table, then build the Mapping Tables from the temp table using Resident loads, then drop the temp table, like so:

TempTable:

SQL SELECT

     key,

     field1,

     field2,

     field3,

     field4,

     field5

from sqltable;

Map1:

Mapping LOAD

     key,

     field1

Resident TempTable;

// --- Repeat for all the other Mapping LOAD statements ---

DROP Table TempTable;

Cheers!

[stevedark]

Hi Mark,

Being a tidy freak that likes code re-use; I would not repeat that mapping statement in that case but use a loop instead:

for i = 1 to 5

     Map$(i):

     MAPPING LOAD
          Key,

           field$(i)

     RESIDENT TempTable;

next

As well as loving ApplyMap I also think dollar sign expansion is pretty cool too. 

Steve

[dvasseur]

Another thing of interest: Mapping Tables cannot be dropped. The DROP statement won't raise an error but the table won't be dropped I had a hard time figuring this out while loading many XLS file in a For each loop.

[m2r4miller]

Absolutely, but the DROP statement in my example was dropping a table - not a mapping table.

[m2r4miller]

That makes great sense! I like to use FOR loops as well. In my example I was just being lazy with field1, field2, etc. I don't expect anyone is actually hitting production tables like that, but then again you never know!

[stevedark]

You are absolutely correct, apart from Address fields 1 - 5 I've never come across any fields that are conveniently enough named to do this.

What I have done though is got a list of fields loaded in an Inline table, enumerated around those to build the Mapping tables (using for i = 0 to NoOfRows('Temp_MappingFields') -1 and then a Peek statement) and also whilst in the loop built a variable with all of the mapping statements in it,doing something like:

let vMapStatements = vMapStatements & 'ApplyMap(' & chr(39) & 'Map_$(vField)' & chr(39) & ', KeyField) as $(vField),'

With a bit of dollar expansion in the load statement you have some nice succinct code and you can add extra fields from the table you are mapping from simply by adding the field name to the Inline table.

Steve

[#10]

Thanks Henric, and others ..

Is there any comment on MAPPING table is size? if the mapping table used in ApplyMap is large if it is over 1M or 100M does it effect anything?

Ersen

[hic]

@ David VASSEUR

You write that "Mapping Tables cannot be dropped". I say: Au contraire! They are automatically dropped at the end of the script run.

HIC

[hic]

@ ebolluca

I have never hit an upper limit on mapping table size, so I assume that the limit is 2 billion distinct values (as the limit is in all other places in QlikView). Which means that if you have a large mapping table, it might consume a lot of memory during the script run, but not after that.

HIC