Tuesday, March 5, 2013

[TABLEAU] Using Aggregated Extractions


               Using Aggregated Extractions

In Tableau, you can reduce the size of your data by taking an extract and choosing the aggregate options. You can aggregate the extract from all data sources. This changes the behavior of the extract.
This example uses the Superstore sample.

Create an aggregated extract

Step 1
On the Dimensions pane, click the menu arrow and select Hide All Unused Fields.



Step 2
Select Data > Extract.

Step 3
In the Extract Data dialog box, under Aggregation, select the check box for Aggregate data for visible dimensions.



Step 4
Click Extract.

Step 5
In the message box, click Show Details to see information about how creating an extract may change the data displayed.

Step 6
Review the information, and then click OK.



Step 7
In the Save As dialog box, give the extraction a descriptive name, and click Save to save it in the default location of My Tableau Repository\Datasources.




Filtering a value

You may also wish to aggregate each record set. For example, filter out values.

Remove a filter
  • In Step 3 above, Tableau will automatically include any global filters in the extract filters.
  • Click on any filters you want removed and then click Remove.

Add a filter
  • Click Add and choose the field. You will be offered the standard filter options for the field.
  • Setup the filter as needed.
These options will filter the data included in the extract. After you create the extract, you will need to recreate the extract to change these values.


Filtering to a time period

You may also wish to aggregate each record set. For example, only show data per month instead of hourly.
  • Check the Aggregate data for visible dimensions
  • Check Roll up dates to.
  • Change the drop down to match the time period you want the data summarized to.
These options will summarize the data included in the extract. After you create the extract, you will need to recreate the extract to change this.



General comments

Number of Records
  • In some cases, the Number of Records field may seem to be missing. In these cases, it is hidden because the original data set did not use it.
  • Number of Records field may not match the original count. For example, if you roll up dates to a summary, the new Number of Records will show the new record total not the previous value.

Calculated fields

Calculated fields dependent on hidden fields may be included the resulting extract and marked as errors. You can delete these fields.



Custom SQL

You can use custom SQL in the original data source. The extract process will use the custom SQL then apply any filters used in the extract process.






No comments:

Post a Comment