Dates incorrectly displayed after Look is downloaded to Excel - Unparseable date

[Stefan2]

Hi,

Date entries are shown as follows, Unparseable date: 'Wed Mar 02 00:00:00 UTC 2022' after I download to Excel. Dates are shown correct when Look in Excel format is sent via email.

This data has been downloaded to Excel before without any issues.

Any help is greatly appreciated.

kind Regards

Stefan

[adit13]

@Stefan2 I am also having the same issue as of yesterday! I expect downloaded time data in the format YYYY-MM-DD but currently seeing it in Day Month DD 00:00:00 UTC YYYY

[DanielJ]

Seeing the same issue. It’s working fine when exporting to CSV, but not XLSX

[Matthew_Burlew]

Same issue here.

[Stefan2]

@adit13, @DanielJ, @Matthew_Burlew 
thanks for your feedback. 

@DanielJ - on exporting to CSV it’s not working as it should either   this is what I am getting:

Thu Feb 17 00:00:00 UTC 2022
Sun Sep 12 00:00:00 UTC 2021

[SarunasH · Accepted solution]

We had the same problem, asked support and got this response (which worked):
“As a workaround for the moment there's a legacy feature you can enable to get the expected output in the download. Could you go to Admin->General->Legacy Features and enable the option `Legacy Query Streaming`”

[Stefan2]

Hi @SarunasH 

Thanks for your feedback. I have enabled the feature in question & it works fine for both XLSX & CSV downloads.

Regards

Stefan

[adit13]

thanks @SarunasH ! This worked for me too.

continued: It looks like Legacy Query Streaming will be removed about a month from now. Is this something Looker will be addressing before the next relevant release or am I individually supposed to resolve the dates-downloaded-to-csv regressions? I ask because I have downloaded the data before and was not having this issue up until only very recently. There are also no transition guidelines here for when Legacy Query Streaming gets removed. I would like to know what the new streamers are doing so I can triage accordingly!

[SarunasH]

@adit13 while discussing with Looker support, I got this information: 
“There's an ongoing issue that our engineering team is looking into that matches the behaviour you're describing. This is an unexpected behaviour we've found in the most recent update to version 22.2 [...]
Our engineers currently estimate they'll have the issue resolved in a couple weeks time, possibly sooner”
I really hope it will be fixed before Legacy Query Streaming gets removed.

[Huth]

We are having the same issue. Legacy Query Streaming fixes the issue but may impact query performance and this legacy feature is being removed. Will Looker fix this soon?

[Maurice_Richard]

Same issue here. Important to fix before legacy feature is removed.

[nico_hein-16299]

Agreed, and therefore raised this with Chat Support. The Product team should be aware by now, still hoping for confirmation here.

Other than the export to .xlsx and .csv, tab separated .txt was not affected in my test and may be a workaround in worst case.

[Chr1s]

continued: It looks like Legacy Query Streaming will be removed about a month from now. Is this something Looker will be addressing before the next relevant release or am I individually supposed to resolve the dates-downloaded-to-csv regressions? I ask because I have downloaded the data before and was not having this issue up until only very recently. There are also no transition guidelines here for when Legacy Query Streaming gets removed. I would like to know what the new streamers are doing so I can triage accordingly!

Following up on adit13’s comment here. According to the documentation it is set to be eliminated in an upcoming release (22.6, just 12 days from now). We need a viable path forward before it is deprecated.

[nico_hein-16299]

I can recommend reaching out to Chat Support to get the relevant information. 

[masutaka]

https://docs.looker.com/relnotes/v22-changelog

FIXED Under certain circumstances, Looker would render dates from Explores and Looks into XLS and CSV incorrectly. This behavior has been corrected.

That may have been fixed. You may want to check with “Legacy Query Streaming” turned off.

[nico_hein-16299]

Thanks for the heads up! Just turned off the legacy feature and it is working again. (Version 22.4.21)

[DanielLee]

Has not this issue been already fixed? And I don’t need to turn on or off “Legacy Query Streaming”?

I’m experiencing this issue again. It works fine when I download files in CSV then open with Excel.

Is it recommended to turn on the “Legacy Query Streaming”?

[cannasafe]

This isn’t fixed. We’re still seeing it on measures.

[StudentOfJazz]

This is still broken for us.

[SarunasH]

The same for us. It was fixed with some version but then with the next one it’s broken again. 

[Yurii_Yashchenk]

I think something goes on terribly wrong with the Looker development cycle. The latest minor 22.6 release broke the patch of this issue from the previous release. The issue is quite severe, but Looker has been fixing it for almost two months. Also, I think using terms like "Legacy" for stable things is confusing, notably with notification that it's going to be removed regarding the default feature being too buggy.

Legacy Query Streaming in 22.6 hast note “This feature will be permanently removed upon upgrade to Looker 22.8, which becomes available about 9 days from now.”, but the actual one is still unstable.

[william_elder-1]

Anyone got an update? I’ve ‘automated’ some Looker reports into Google Sheets but they require date formatting each time - hardly automated now, it’s still requiring my time to modify the date format. 

Please let me know when this fix will be in place. 

[Stefan2]

Hi,

This is not solved at all;

• Excel format - I get Unparseable date
• CSV - its correct BUT instead of ‘€’, I get the following symbols ‘€’, which creates another problem ?

can someone look into this please?

Regards

Stefan

[djdevin]

The “Best answer” no longer works as the legacy feature was removed. This issue should be un-marked as “Solved” as the new feature is broken.

Currently there’s no way to correctly export dates in Looker which seems pretty critical to me...

[Stefan2]

agreed @djdevin and thanks