I am trying to optimise a Qlik Sense App, but I can't explain the size of the app.
To be clear, I have two apps, a "small one" and a "big one", and I did the following comparison : -Small one 9,5 millions lines, big one 11 millions lines. -Small one 74 fields, big one 84 fields. -Small one 8 sheets, big one 9 sheets. -Small one 5 master items, big one 21 master items.
So, it is obvious that the size app should be bigger for the big one. But the file size that can be seen from the hub is 166 Mo for the small one, and 602 Mo for the big one.
I can't explain such a difference with the comparison I made, since the big one doesn't have 3 times the number of lines, same for the fields, etc. Is there something I'm missing ?
You should also keep in mind other things, let's start from the first one: the values of your fields.
I make an easy example: if you have a field containing 1 milion lines but every line is one-digit (ex. A,B,C..) it will be lighter than a field containing 20 digit per line. The example is more realistic with keys in your data models constructed by two or more fields aggregated and for which is better to use the autonumber() function
Let's now talk about the data model. Implementing a star schema model will reduce the weight of your application instead having a unique table.
Finally, in front end there are sheets and sheets depending on the amount of objects used and the density of measure/expressions.
Please, feel free to leave a like if this comment helps you out!
The storage is in general not linear - neither from the vertical point of view (number of records) nor to the horizontal one (number of fields).
A single field with larger string-content or a quite granular timestamp-field (maybe even with a row-level formatting) may make the difference and becoming larger in size as all other data together.
As others have pointed out, size is related to the uniqueness of the data, not so much the number of rows or number of fields.
In the QMC are you looking at the "File Size" or the "Base memory size"? File size is size on disk, not so useful. It's the "base memory size" column you want to look at.
Here's a post that explains a bit about App data size and techniques for reducing size when necessary.
You should also keep in mind other things, let's start from the first one: the values of your fields.
I make an easy example: if you have a field containing 1 milion lines but every line is one-digit (ex. A,B,C..) it will be lighter than a field containing 20 digit per line. The example is more realistic with keys in your data models constructed by two or more fields aggregated and for which is better to use the autonumber() function
Let's now talk about the data model. Implementing a star schema model will reduce the weight of your application instead having a unique table.
Finally, in front end there are sheets and sheets depending on the amount of objects used and the density of measure/expressions.
Please, feel free to leave a like if this comment helps you out!
The storage is in general not linear - neither from the vertical point of view (number of records) nor to the horizontal one (number of fields).
A single field with larger string-content or a quite granular timestamp-field (maybe even with a row-level formatting) may make the difference and becoming larger in size as all other data together.
As others have pointed out, size is related to the uniqueness of the data, not so much the number of rows or number of fields.
In the QMC are you looking at the "File Size" or the "Base memory size"? File size is size on disk, not so useful. It's the "base memory size" column you want to look at.
Here's a post that explains a bit about App data size and techniques for reducing size when necessary.
