piled bar graphs not possible?

[chriscools]

Hello,

can someone help me how i should make a graph like the one below?

i would think that it was an option under the "bar graphs" where i could say that the different dimensions are piled.

what i want is the % part of sales per product family.

and i want 1 bar to show sales for the current selected month, and 1 bar that shows the % division for whole the year.

thanx!

chris

[swuehl]

Chris, maybe like attached?

[chriscools]

Thank you!

i'm trying this right now.

but isn't it necessary that the KPI table has a link somewhere to  the 'facts'?

because i see that you two unlinked tables, what's the theory behind this?

thanx!

chris

[MayilVahanan]

HI

You can use like this also,

Dimension

=ValueList('Share of Month','Share of Year')

ProductCategory

Expression

if(ValueList('Share of Month','Share of Year') = 'Share of Month',

sum(Value) / sum(total Value), sum({<Month = {"<=$(=max(Month))"}>} Value) /sum({<Month = {"<=$(=max(Month))"}>} total Value)

)

Mr.swuehl used KPI value in inline table, for dimension purpose..

[chriscools]

is there a reason to have a preference for the one or the other solution?

i will try the 2.

thanx!

chris

[swuehl · Accepted solution]

"is there a reason to have a preference for the one or the other solution?"

One is script based, one is front end only.

"but isn't it necessary that the KPI table has a link somewhere to  the 'facts'?

because i see that you two unlinked tables, what's the theory behind this?"

No, it it's not necessary. If you don't have a link, your measures won't be grouped by your dimension, i.e. all records in your selection are inputted to your expression for both dimension values.

[chriscools]

thanx, for the explanation, i got the both ways to work!

grtz,

chris

[Or]

If you don't mind me asking - *why* do you need this particular graph? Wouldn't you be better off with a pair of non-stacked bar charts, or a grouped bar chart with Product Category as the dimension?

[chriscools]

Hello,

if you look at my first post in this question, there is the graph i had in my excel reporting.

and i find it visually easier to compare the size of a product category if they are right next to each other.

in excel i could include the % values, and that i can't find in qlikview,

and that indeed makes it harder to compare.

but still i find it easier to compare than in the grouped bar chart where you need to allways jump over a few bars.

but tomorrow i might try the pair of non-stacked bar charts, to see what that does visually to compare.

grtz,

chris

[Or]

Let me know how that works out, if you don't mind.

Personally, I'm not a proponent of stacked graphs - they're very hard to read because color chunks don't start at the same spot. In your example, the 14% and 12% dark blue chunks look like they're about the same size to me. Insofar as placing the actual numbers in the graph - sure, you can - but then you're better off with a graph-and-table combination.

Assuming you have the space, here is how I'd probably go about doing this:

Note:

* We can easily compare categories across periods by using the left graph. This is particularly nice because we don't have to use the legend to figure out which category is which.

* We can easily compare the makeup of each period using the right graph - now the difference in the fourth category (yellow) is a lot easier to see even though we're still looking at periods rather than categories.

* We can easily get the numeric values using the table (the pivot table's dimension can be swapped by dragging, allowing easy comparison of either periods or categories).

Ultimately, it's a matter of personal preference, of course - mine is to use less information and more graphs rather than forcing a single graph to allow multiple (and significantly different) types of analysis.