Data Studio
3.9K members online now
3.9K members online now
Ask questions about Data Studio
Guide Me
star_border
Reply

Non-ASCII dimension labels (e.g. German umlauts ä,ö,ü,ß) cause issues in reports

Visitor ✭ ✭ ✭
# 1
Visitor ✭ ✭ ✭

In Google Data Studio Beta, I have created two similar reports. The first report is based on a BigQuery table (loaded from Cloud Datastore backup), but it only shows random reference IDs for a "category" dimension, because the strings are stored in a second (category) table. I would like the users to see the actual titles of the categories, rather than their category ID. I will refer to this string field as "CategoryTitle" from now on.

 

So I have created a new version of the same report, this time based on a custom query over the same table joining the second table to include the "CategoryTitle" of each referenced category.

 

On the first look, the new report works as expected. I can use the strings as dimension values. However, if the string of a dimension value contains a German umlaut (e.g. ä ö ü ß), the related metric values don't appear in some charts. The metric values should be there (a simple function on a field as sum) and are shown in other charts and controls.

 

Bar Chart:

  • If I only use one dimension (with some of the "CategoryTitle" values having umlauts), the bar chart shows all bars (as expected).
  • If I add a second dimension (a number field), all bars will disappear where the CategoryTitle contains an umlaut, the bars of ASCII-only category titles are not affected; I would expect to see all bar charts, including the bars of category titles with non-ASCII characters.

 

Bar Chart with Filter Control:

On another page, I have a filter control that uses CategoryTitle as dimension and again the same field as metric (the metric is also shown in the filter control to see the sum per category title). The bar chart only uses the category title as dimension and again the same metric field.

  • If in filter control, all category titles are selected, all bars' metric values sum up to the sum of all metric values in the filter control (as expected).
  • If I deselect just one of the category titles with umlaut, the sums in the bar chart are reduced by the according value (as expected).
  • If I only select categories with umlaut, the sums in the bar chart match those of the selected dimensions in the filter (as expected).
  • If I deselect all but one of the categories with ASCII-only characters, the sums in the bar chart match those of the selected dimension in the filter (as expected).
  • If I deselect all but one of the category titles with umlaut, no bars are shown at all, but the y-axis scales to a high value as if all categories had been selected; I would expect to see bars where the sum matches the metric value shown in the filter control for the selected category title.

 

Using CategoryTitle as labels in a pie chart seems to work as expected. So the issue might not affect all charts.

 

I also have used the category table as data source for a third report, again using the CategoryTitle as dimension. I could not reproduce the issues with this report. So the issue might affect only reports that have a custom query (BigQuery?) as data source.

 

I can share screenshots of the reports and code samples if needed.

Non-ASCII dimension labels (e.g. German umlauts ä,ö,ü,ß) cause issues in reports

Visitor ✭ ✭ ✭
# 2
Visitor ✭ ✭ ✭

Hi Anastasios!

 

I just made a similar post on this. I've experienced the exact same issue. 

 

Did you find any solution to this? 

 

Cheers,

Isak

 

Non-ASCII dimension labels (e.g. German umlauts ä,ö,ü,ß) cause issues in reports

Visitor ✭ ✭ ✭
# 3
Visitor ✭ ✭ ✭

Hey,

 

Okay, I just found a (very weird) solution to this problem.

 

Create a calculated field with the following function: 

REGEXP_REPLACE(field_expression, regexp, replacement)

 

1. replace field_expression with the dimension containing the unreadable characters.

2. Now to the weird part: put whatever character you want - e.g. "a" - instead of regexp and replacement.

3. The result should look like: REGEXP_REPLACE(YOURDIMENSION, "a", "a")

4. Use your new calculated field in the filter instead of your previous dimension.

 

I hope it work for you as it did for me...

Non-ASCII dimension labels (e.g. German umlauts ä,ö,ü,ß) cause issues in reports

Visitor ✭ ✭ ✭
# 4
Visitor ✭ ✭ ✭

Hi Isak,

 

both issues that I have described in my post had been resolved a few months later (maybe in August 2017?). I haven't done anything on my part, so I assume they were fixed with a new release of Data Studio around that time.

 

Please note that my issues occurred when joining tables from BigQuery. If I understand your other post correctly, you experience the same issues when joining tables from MySQL. I assume different code is used internally to read from BigQuery and MySQL and perhaps the bugs have been fixed only for BigQuery sources, not for MySQL.

 

Cheers,

Ani