Loading data in Incrementally loading cube with deleted ecdata files.

This issue is regarding trying to load data on a cube under specific conditions-

1) some tables of the cube have been set up to load data incrementally.
2) you have had to manually delete backup files due to lack of disc space.

In this case no matter which load you try to do- schema changes, load full data, delete existing data and load sample, cube will fail to load.

Work around:
copy the cube in another name and run it again and change the cube in dashboard.

Soluution:
I found no solution to above problem. It continuously fails with message: "failed to load cube". No matter what you try, it will continue to show the same message.
The reason for this seems to be that in some procedure it has stored that there is an existing copy of the cube data (ecdata file). However upon not finding it, instead of letting you do a full load, it keeps throwing error.

Time based transformations in Sisense.

When trying to implement time based transformations sisense offers some good options, if that is, your time requirements are limited to month, quarter or year.
If you want to implement last week, 28 days, last 30 days average/sum then you can run into some big confusion.

In other tools this is managed by transformations, however since no such feature exists in Sisense,  one will have to implement it using custom tables. However upon googling you will come across two very helpful features-

RPAVG
and
RPSUM

These are supposed to give you running period average and running period sum.
http://pages.sisense.com/rs/sisense/images/SisenseFormulaReference.pdf

However, if you try to implement it you will see some unexpected behaviour.
Upon contacting sisennse you will come to know that the features are no more supported, despite being mentioned on their documents.

Recommended solution-
Please implement your requirement through another SQL-based query within your E-cube. This will be additional overhead for maintenance and cube build time. 

Why not to use views in Sisense

Sisense works basically on importing all of your data into its Ecubes and then doing the processing which you require. It insists that it can act as your ETL provider too. But there are some issues which novices can be hit hard by. One of them is using views.
By using a view you have basically pushed the logic at your warehouse instead of at sisense. When you try to import a large table into Sisense its structure gets there within no time and then on, it will only consume your time when its time to load data. However when you try to upload a view with a large resultant volume, you will observe that it takes a lot of time to get the structure in your Ecube.
It can be irritating to sit in front of the idle screen for long and sadly there is no way to cancel it, pressing any buttons will crash the machine.

Exporting/importing Ecubes.

The first thing to remember is the difference between Ecube data and ecube structure.

Its the difference between .ecdata and .ecube data files- while the former contains the data from the last build of your cube, the latter is the structure of the cube you have created.

The second thing to remember is how to export. Based on your security settings there can be multiple ways to do this.  Data file- .ecdata are generally large and may require you to setup an ftp; however you can simply copy the .ecube files from one server and paste it on another.

Now you can open them from the sisense elasticube manager. Though you will have to build it again as its only structure and no data.