I use custom events for tracking statistics about some deprecated modules are used by users. And I`ll want to remove migrations from deprecated module to a new one when amount of usages will be lower a "waterline".
So, it is not enough convenient to track it via clicking on a date on a graph and check amount of events at the date. Could I somehow switch a type of values on a graph to absolute values?
Mike from Fabric here. For the graphs, we will either show the percentage if the custom attribute is a string or the 25th, median and 75th percentiles if the custom attribute is a number. However, the top 10 custom attribute count will be present below the graph.
Related
I've been dealing with this issue in various ways throughout my time with this dataset in Tableau.
As you can see, the Total count of properties for each city is including properties that have been successfully filtered out of view. Why? The dyn.RANKED Profitable Investments (grouped) variable on the Filter shelf is an attempt to double down on the same as the first line of the Calculated Field - to ignore the unwanted properties in each city. The view ignores them, but the totals do not.
If the Watershed Property pill is removed from the Rows shelf, then the dyn.NumProps_in_City results shown on the table are each the same as the Totals you see here (i.e., despite the first line of the calculated field, properties that do not meet that opening condition are being counted)...despite the view with the Watershed pill knowing not to show them.
Also if the Watershed Property pill is removed from the Rows shelf, then the dyn.RANKED Profitable Investments (grouped) variable on the Filter shelf suddenly only has one category to choose from (i.e., 'INVEST') if you go to edit the filter. Which would be great since that's the category I care about, but not if the counts are including things that are not in that category despite the filter.
Messing around with Include, Exclude, and Fixed in the calculated field doesn't seem to work here since I can't figure out how to get around various aggregate/non-aggregate and/or ATTR errors no matter where I place them. Plus, my incorrect counts are not suffering from an LOD issue - the LOD is correct - it's an issue of not consistently filtering out the unwanted rows at the desired LOD.
Please advise!
Thanks,
Christian
It seems that dyn.Ranked calculated field calculates the value prior to filtering. This may happen if you have used any LOD calculations in the syntax.
Simply right click such fields on filters shelf and click add to context. This will cause LOD calculations to calculate after the filtering.
see this link, the context filters are above the LOD calculations, in order of precedence; but measure filters are below the LOD calcs. Therefore if measures are used as filters, these have to be added to context so that their order of precedence is above such calculations
I am creating a Dashboard using as a visualization a heat map. Everything was OK until I changed the parameters of my metric, the chart disappeared and I've got this message: 'Filter excludes all data'
The only modification that I've done is to set the Include Distinct Elements to true within the Count Parameter option of the metric.
What could be happening?. Do I need to set another parameter to get the count of distinct elements that I need?
Regards.
surely the metric is level with some attribute that is not inside the visualization, if the filter has date for example, include it in the visualization.
"Filter Excludes All data" is a default warning message that you will get in MicroStrategy when a reports/visuals/dashboards does not return any data.
https://community.microstrategy.com/s/article/KB47557-How-to-Properly-Suppress-the-Message-Filter-excludes-all?r=1&Component.reportDeprecationUsages=1&Headline.getInitData=1&ArticleView.getArticleHeaderDetail=1&Quarterback.validateRoute=1&RecordGvp.getRecord=1&ArticleRichContent.getArticleAuthor=1&ArticleTopicList.getTopics=1&ArticleRichContent.hasArticleAccess=1&ForceCommunityFeed.getModel=1&ArticleRichContent.getTopicsAssigned=1
There are n number of reasons for a report which cannot return data, please check the following steps to debug the issue,
As per your first image it shows a date range as filter, after you have changed to "Include only distinct" this date range might get affected so please put the objects in grid and apply the date filter and check whether it returns the correct data.
If it returns, check whether all the metrics returns a value.
Check whether candidateID, date and people attributes/metrics are properly related.
These steps will show, where the problem is, still if you could not figure it out export the dashboard and share it with MicroStrategy Tech Support team for debugging.
Hope it helps.
To understand it fully I will give a clear example:
Our webshop sells zip ties. Every zip ties can be tightened between a minimum and a maximum diameter. For example say from 0.5" to 2".
In most of the cases the buyer needs a zip tie for a specific needed diameter. What we need is to have a possibility to get an input for the needed diameter and list all the zip ties where this input is between the minimum and maximum diameter.
I have tried different variations with layered navigation but it can not be done with it.
Any help is welcome.
You need to customize blocklayered module, some tips:
let's imagine in shop each product have 2 different features "min diameter" and "max diameter" and you already created filter for them, checkboxes.
hide these blocks by display:none
add in template input field for the actual diameter value.
after customer's input, select all min/max checkboxes that satisfy the condition
call blocklayered method to send the filters.
I am working with Endeca 6.4.1 and have many auto-generated dimensions present in my pipeline (mapped using Dev-studio), the application's indexing is CAS-less. So only FCM is creating Dimensions and assigning dValIds. I am using Endeca SEO, so the dVal Id directly reflects in my URL, and if an auto-gen dimension's value's Id changes, a link to that navigation State is lost.
I have a flat file as the dimension's source, for example
product.feature|neon finish
What I want is that, if the value some day changes to Neon-finish or Neon color, the dValId that was assigned to neon finish should be transferred to the new value. I can keep a custom mapping of the change to track that neon finish has been changed to a new value.
Is there any way to achieve this, may be by using some manipulators?
Please share your thoughts.
There are two basic ways to do this:
1) Update the state files when you change a dimension value (APPDIR/data/state/autogen_dimensions.xml ). This would most likely be a manual process.
2) A more robust but complex solution is to change the dimension values to be some ID number and use a synonym for the display name. Then the display name can change without a change to the id number. This may require some serious changes to your pipeline.
Good luck
I am looking to create and use a gauge chart, but I am trying to figure how to compare selected vs the complete universe. For example, let's say I want to measure a monthly spending average of one department vs. the spending average of the entire organization. How do I make the organization average the segment marker? or make it a static variable that does not change with the selection?
As is the case with a lot of questions that pop up here, you can use set analysis to achieve this. In you gauge expression you could do something like this:
Assuming you are going to select one or more departments in the current state, you could do
percent_of_total_spending = avg({$}spending)/avg({1}spending)
The {} syntax is used in conjunction with the set identifiers $ and 1 to denote the the set of the current selections and the entire universe respectively. So the above expression will give you the percentage of average organizational spending for whatever departments you have selected.
You can use set modifiers in addition to the set identifiers to further customize what population you are including. See here for more information.
All you need to do is use set analysis and tell the expression what to ignore in the selections box.
Example:
Sum({<Month=, Year=>}Spending)
What the above is telling you is that it's taking the "Month" field and setting the selection for that specific expression to nothing, as well as the "Year" field. If you were to do the following:
Sum({<Month=, Year={"2014"}>}Spending)
It will calculate for that specific year, or even this
Sum({<Month=, Year={"2013","2014"}>}Spending)
Will make it calculate for both 2013 and 2014. Read up on the helpfile as well by pushing F1 and searching for set analysis.
So something like this will work for you
Sum(Spending) / Sum({<Month=, Year=>}Spending)