The Hypercube Results are too large Error for Polygon Maps

[#1]

Hi All,

We are evaluating Qlik Sense for the maps requirements and I am stuck with this error " The Hypercube Results are too large" Error for Polygon Maps. No matter what I do -  I get  this error. I am trying to make polygon maps for Australia Postcodes and so far no luck. So here is what I have in my data load editor

Geo:

LOAD

    "Australian Postcodes.Name" as Postcode,

    "Australian Postcodes.Point",

    "Australian Postcodes.Area"

FROM [lib://AusPost/Australian Postcodes.kml]

(kml, Table is [Australian Postcodes/Postcodes_Simplified]);

LOAD

    @1 as Postcode,

    @2 as Suburb

FROM [lib://AusPost/AU_SuburbName.csv]

(txt, codepage is 1252, no labels, delimiter is ',', msq);

GeoAggr1:

load   

GeoAggrGeometry("Australian Postcodes.Area") as Combined

resident Geo

group by Postcode;

Data loads fine but when I try to build the map I get this error.

When it asks me to select the dimension I select - "Postcode" and then the next step it asks me to "select a geographical field" then I select "Australian Postcodes.Area (area)"  and I immediately get the error.

I even tried "Suburb" and "Australian Postcodes.Area (area)"  and I immediately get the error.

I will really appreciate if someone can help me with this issue. This is a showstopper for my project and the client is ready to consider other tools look Tableau.

I even tried to reduce the KML file size by reducing the precision but after the the reduction, Qlik sense doesn't recognize column names of the KML file.( http://www.gearthblog.com/blog/archives/2016/03/making-kml-files-smaller-reducing-precision.html )

I am attaching both the files for your reference.

I am attaching a 3rd file as that is the one I have to use ultimately make the polygon maps ( third file name is - P_M_R_Final.csv)

[Anonymous]

Hi Ajay,

You found a solution? I've download your files and tried to load your map and it works for me.

First of all, I reduced the polygons to one decimal. Then I load the table with the zipcodes and areas doubling the PostCode (kk.Name as PostCode and kk.Name as PostCodeDesc). Finally I load the map using the Area field and the doubled field as a dimension for the tooltip.

I recentlly have a similar issue, if you want more information you can check this link Map Object don't load all areas

GIS:

LOAD

    kk.Name as PostCode,

    kk.Name as PostCodeDesc,

    kk.Point,

    kk.Area

FROM [lib://AU_KML/kk.kml]

(kml, Table is [kk/kk]);

Master:

LOAD

    PostCode,

    "Suburb Name"

FROM [lib://AU_KML/AU_SuburbName.csv]

(txt, codepage is 1252, embedded labels, delimiter is ',', msq);

Facts:

LOAD

    STATE,

    Postcode,

    "Month",

    article_count,

    sl_total,

    ll_total,

    sp_total,

    sl_minus_um,

    ll_minus_um,

    Mat_group,

    Num_tran,

    Quant_Sum,

    Amt_sum,

    num_stores,

    POPULATION,

    PERCENTILE

FROM [lib://AU_KML/P_M_R_Final.csv]

(txt, codepage is 1252, embedded labels, delimiter is ',', msq);

In the map add Area and PostCodeDesc.

[#3]

Thanks Jordi,

I will try this.

[#4]

Hi,

Australia has some postcode areas that are quite large. So even if you manage to load them they might be slow to display since the areas contain so many points. In IdevioMaps (se http://bi.idevio.com) we provide postcode areas for many countries including Australia. We have simplified them so that they are useful and also provide aggregated levels with less digits so that it is easy to build drill-down hierarchies.

Just drop us a message at info@idevio.com to get an evaluation license.

Btw, the thread https://community.qlik.com/message/1052942?et=watches.email.thread#1052942 is very similar.

/Johan

[#5]

Hi Jordi - can you tell me how you were able to reduce the polygons to one decimal? Was there a tool that you used?

Kind regards,

Nicole

[#6]

Nicole

You can use these steps - https://community.qlik.com/docs/DOC-8969 . Go to the section Simplify the KML and you will be able to reduce the size of the KML.

Simplify the KML

For our example file, we have gone from a 9MB Shape file to a 20MB KML. This is going to cause Qlik Sense to lag a bit when in operation. But why is it doing this? Essentially the level of detail in our KML is far too high. As an example, look at this screenshot of our output KML. I don't think many people care which county that particular rock is in!

We need to simplify our KML which can be done while converting it from the original Shape file with the addition of the "-simplify" flag to our ogr2ogr command

1. Re-run the ogr2ogr script with the simplify flag applied

• e.g. ogr2ogr -f "KML" -simplify "10" county_region.kml county_region.shp

2. Review the newly created KML file to make sure the file size and accuracy are still fit for your purpose. Adjust the simplify factor as required

As some examples of the different levels here are some sample outputs

[#7]

Hi Ajay, thank you so much for the response! Question - my starting point is a KML file that is very detailed (225MB, 13 decimal places). I do not currently have shape files, but am using this zip code level KML to aggregate to my geography hierarchy, and the detail is still far to great for Sense to handle. Do you have any suggestions on reducing the precision of an already existing KML file?

Since I am using US zip codes, I will try to see if there are any available shape files so I can try out the tool you suggested.

Thank you again for your help!