Hi Cynthia, well...I've been thinking about it, but have found a bit of problems using a stored process linked to an information map in EG (in other words: it doesn't work smoothly as it does in Web Report Studio...it doesn't call the stored process before the information map in eg). ... View more

Hi, here is the code using the libname: /////////// libname TEST infomaps user='dtiles' pw='******' metaserver='icodevsas02' mappath='Shared Data/StatSS' preserve_tab_names=yes; options validvarname=any; proc contents data=test.famiglie_es; run; ////////// This is the result: /////// SAS 14:30 Saturday, February 13, 2010 1 La procedura CONTENTS Nome data set TEST.famiglie_es Osservazioni . Tipo elemento DATA Variabili 9 Engine INFOMAPS Indici 0 Data creazione . Lunghezza osservazione 0 Data ultima modifica . Osservazioni eliminate 0 Protezione Compresso NO Tipo data set Ordinato NO Etichetta Rappresentazione dei dati Predefinito Codifica Default Elenco alfabetico di variabili e attributi N. Variabile Tipo Lungh Etichetta 8 Anno Alfanum 32 Anno 1 Istat Alfanum 200 ISTAT 2 Istat2 Alfanum 32 Comune 3 Istat2_2 Alfanum 32 Unione 4 Istat2_3 Alfanum 32 Zona Altimetrica 9 N__famiglie_unipersonali Num 8 N° famiglie unipersonali 7 Numcomponenti Alfanum 2 Numero componenti 5 Numero_Di_Famiglie_Unipersonali Num 8 Numero famiglie 6 Residenti_In_Famiglie_Unipersona Num 8 Numero residenti ////// In the information map the variables are inserted into a folder, but I don't see the name as shown on the doc (folder/name of the data item). The other code is this one: ///// ods trace on /listing; proc infomaps metauser="dtiles" metapass="**********" metaserver="icodevsas02" ; open infomap "famiglie_es" mappath="/Shared Data/StatSS"; list; run; ods trace off; ///// In this case, I don't get the dataset that I should use with ods output... ... View more

That's would be wonderful, however I can't find the correct table (I've tried to follow this paper http://www2.sas.com/proceedings/forum2008/086-2008.pdf, but I couldn't find the ods table for the proc infomap, using ods trace on, even with listing). ... View more

Yep, I've already read it and I'm using it...but the proc contents it's not sufficient for what I'm looking for (for example, it doesn't show into which folder an item is contained...while proc infomaps does)...I was just looking for a way to output the "list infomap dataitems" output in a dataset, in a manner similar of the proc contents... ... View more

Hi Cynthia, thank you very much for your answer! Finally, everything is explained :). This was of course a simplified problem: the real problem was that the format is dynamically updated every night, and so we are sure that there aren't any cities with code '999999' (our total format), but...maybe tonight the city 199 is inside a format, tomorrow it's outside any category and then it should fall back to the "other" category As we've got a complete list of the cities, we've done a simple code which associates each code of a city who doesn't fall into a format created by the user into a new "None" format, and we have created the format notsorted, ordered as we wanted. This is unfortunately less elegant than using the other format (which I found brilliant, until I was asked this particular order...), but it's still the best solution we could find. Again, thanks for unveiling the problem 🙂 ... View more

Hi to all, in the end we solved our problem by changing our format and avoiding using the "other" option in the formats...but still, I'd like to understand what I'm missing here. Consider this sample code: ////////////////////////////////////////////////////////////////////////// proc format cntlout=work.fmt; value $first (notsorted) '1' = 'First' other = 'None' '999' = 'Total'; value $second (notsorted) '2' = 'Second' other = 'None' '999' = 'Total'; run; data test; length first $8.; length second $8.; length num 8.; input first second num @@; cards; 4 4 20 5 6 10 1 2 10 1 2 10 3 3 40 999 999 90 ; proc tabulate data=test; class first / preloadfmt order=data; class second / preloadfmt order=data; var num; table first='', second=''*num=''*sum=''; format first $first.; format second $second.; run; ////////////////////////////////////////////////////////////////////////// What I expect (or would like) is a cross-table like thi one: _______| Second | None | Total First | | | None | | | Total | | | Instead, what I get is this: _______| Second | Total | None First | | | Total | | | None | | | What I'm doing wrong, missing? Thanks Daniele ... View more

Hi to all, thank you for your answer! I've found to be part of the problem the use of the preloadfmt, and I've solved 90% of the problems found. Unfortunately, I cannot post the code...it's too much complicated and confused by macro code, and it would be just an headache for each one who would try to help me :). Cynthia, your code looks quite interesting, but I've to ask something about the "notsorted" option: actually, we've got a program running each night which updates the formats. However, we update the format using the cntlin option: we have prepared a table (downloaded from a SQL Server) which contains the minimum fields for creating a format...which field should I had for the "notsorted"? Thanks Daniele ... View more

Hi to all, for the first problem, I've solved by proc tabulating first with the format, and outputting the results in a dataset, that I've used with the proc gchart. In this way I've got one observation for class, and it's easy to order it as I want. For the second point...yes Chris, I mean the axis1 legend... Thanks Daniele Edit: no, I was wrong, I meant the label 🙂 Message was edited by: Daniele Tiles ... View more

Hi Flip, maybe I'm using wrongly the origin option, but it doesn't only move the legend, but it also stretches the diagram...at the point that it doesn't put the legend inside the graph 😞 ... View more

Hi Dan, I'm not sure of what you have explained on your post...how I should use midpoints? Let me explain my dataset: I've got values for all the range of different ages from 0 to over 80 years, and I've to plot the values on a hbar diagram, where: 1) I've applied a format for the age, which groups the ages in 5 years segments (0-4 years, over 80 years old, and so on) 2) I'd like to show this diagram in this order: 80 years and over old 75 to 79 old ... ... 0 to 4 years old 3) This legend (which are the ages formatted) should be pun in the center of the graph, where tha value is 0 (the fact is that I've got also negative values, and the legend is at the left of the negative). Thanks Daniele ... View more

Dear Bill, thank you for your kind answer, but I believe I didn't express myself correctly: what I want is the legend of the vertical axes in the center of the graph (normally is on the left, since this is a hbar graph). Is it possible? Thanks Daniele ... View more