Quicksight dividing latitude and longitude values by 1000

Question & Answer
, , ,
1

I am trying to create a “Points on map” visual in my Quicksight dashboard. The latitude and longitude values appear correct in the dataset, however when I try to use the fields in my map I see it has divided the original value by 1000. For example, a latitude of 40.8422403 in the dataset shows up as a value of .04 (which I’ve checked by viewing the latitude/longitude values in a table visual). The fields are correctly tagged as Latitude and Longitude in the dataset. This is making it impossible to create a correct map. Any idea what is happening here?

2 Likes
2

Hi @margaweb
I had the same issue and could figure it out.
So I created new fields multiplied with 1000.

Strange thing, it doesn’t appear an all dataset.

Let’s see what others say.

BR

5 Likes
3

We had the they problem with latitude and longitude values when we used a PostgreSQL database as source.
Casting the float values as decimal solved the issue for us.

1 Like
4

Thanks this worked for me! Created a calculated fields in the dataset where I multiplied the lat/long fields with 1000 then set their data type to Lat and Long, then used these in the analysis. This appears to be a recent bug, I have created maps before without this issue.

5

@margaweb Thanks for posting your question. If @ErikG’s reply is the solution to your question, could you mark it as “solution” by clicking the check box under his reply?

Thanks!

6

Hi Kristin, will do! However I’d consider this more of a workaround than a solution - it seems like this is a Quicksight bug that others are experiencing. Is there some way to report it?

7

Agree with @margaweb . We ran into this same issue and had to create new calculated fields in the SQL query. This seems like more of a work around than a fix. @Kristin , do you know if this is on the roadmap to be fixed. If it is fixed, we will need to revert the work around as all lat and long fields will be off by a factor of 1000, messing up all of the location data.

8

Hello,
I have encountered the same bug.

I’m querying parquet file using Athena, when I use in-query it doesn’t downscales Latitude and Longitude.
However when I had to import it into SPICE (because of cross-source join) this bug started appearing.

9

This solved an issue in my case as well. Same problem and I had a Aurora based dataset.