Problem to connect from ODate in Power BI(service)

[YarSolntsev]

Hello Guys!

The PBI Desktop did a project with a connection from the data source OData following:

http://192.168.1.32/IdealShop_Test/odata/standard.odata/Catalog_Enterprise?$select=Ref_Key,Descripti...

It’s not all columns in this table!

Everything works fine. The columns "Ref_Key" and "Description" is imported from the table and loaded into the model PBI Desktop. 

However, after the publication it in the PBI and destination data source through Odate, doubt that the set of columns specified after the SELECT, the table is ignored and will always try to boot from all existing columns.

Is this true?

[v-yuezhe-msft]

Hi @YarSolntsev,

Based on your description, you add $select=Ref_Key,Descripti... in your OData feed URL, do you want to get specific columns of table into Power BI Desktop using this method? If that is the case, you can import the whole table into Power BI Desktop, then remove unwanted columns in Query Editor, please check the following example. To remove all unwanted columns, you can select them then click “Remove Columns”, or you can select essential columns(Ref_Key, Description, etc)  then click “Remove Other Columns”.



This way, once you publish the Power BI Desktop, everything should be works well.


Thanks,
Lydia Zhang

[YarSolntsev]

Hi Lidya!

Yes, I can to select and add specific columns of (Odata) Tabe to PBI Desktop by PARAMETR - "$SELECT".

This metod to work is good in PBI Desktop application, but it don't work in PBI. I don't  undestand, probably I need to use pay version PBI? )
In your solved - "Add all columns then remove all unwanted columns" is not good, because UNWANTED COLUMNS to many.

[v-yuezhe-msft]

Hi @YarSolntsev,

As mentioned in my first reply, you can can select essential columns(Ref_Key, Description, etc)  then click “Remove Other Columns” to remove unwanted columns, it will not take you much time if you only want to reserve few columns.

In addition, could you please post the full error message about timeout?

Thanks,
Lydia Zhang

[YarSolntsev]

Hi!
Yes of course.

Any columns name is cyrilic.

I get this error by criate "datasource" for PBI.

Table: "AccumulationRegister_Продажи_RecordType"
Columns: Recorder,Period,Номенклатура_Key,Подразделение_Key,Стоимость,Recorder_Type

Unable to connect: We encountered an error while trying to connect to http://192.168.*.*/IdealShop_Test/odata/standard.odata/AccumulationRegister_Продажи_RecordType?$sele.... Details: ""Hide details

Activity ID:	92f6131c-f3f0-02ac-c6e8-86869d3fb7b0
Request ID:	6f0ddb81-ba81-d5cf-a141-de263dec5948
Cluster URI:	https://wabi-west-europe-redirect.analysis.windows.net
Status code:	400
Error Code:	DM_GWPipeline_Gateway_CancellationOrTimeoutError
Time:	Tue Nov 22 2016 12:11:11 GMT+0300 (Kaliningrad Standard Time)
Version:	13.0.1700.631

[v-yuezhe-msft]

Hi @YarSolntsev,

Large dataset and complex complicated data model could cause the refresh to timeout. Try to reduce the complexity of your report and data model, also consider to break your dataset into smaller datasets.

Thanks,
Lydia Zhang

[YarSolntsev]

Hi Lydia!

As I undestand that problem is realy.
The connect DB by ODate method  between a PBI and PBI Desktop is defferent. (
For PBI Desktop is good work For PBI and the same time it don't work for PBI. Why I should break my dateset? )
How can I break my dateset?
Is to change my dateset from one query -

...&$select=column1,column2,column3,column4...

to both querys

...&$select=column1,column2...

...&$select=column3,column4...

Is't right?

[v-yuezhe-msft]

Hi @YarSolntsev,

Refresh in Power BI Service could be timeout because of large dataset and complex data model.

How many visuals in your PBIX file? You would need to delete extra rows and columns in your PBIX file, and simplify visuals in the PBIX file.

If the above error still occurs, please consider to seperate data into different PBIX file.

Thanks,
Lydia Zhang

[YarSolntsev]

Hi!
A few days a go we made the following. We changed Gateway from virtual machine to real machine. Then our problem was solved! )
From this case we have quastion. Is Gateway mast to installed on the real computer?

[Anonymous]

192.168.xxx is an address on your internal network - we can't see it.  If you copy-paste more details of the query you are running and the output you get, it will  help sort out what's happening.

[YarSolntsev]

Hi Steve!

My more advanced problem in the following.

I try to connect Table to PBI by local GateWay.
If Table is small (2-4 colums and 50000 rows) that everything good connect.
If Table is big (15-50 columns and more 100000 rows) that is can't connect by timeout GateWay.