About RoyPardee

RoyPardee · ‎02-23-2024

sure fair enough. This code illustrates what I just went through. data cyl_attrs ; retain id "cylinders" linepattern "solid" ; input value 2.0 linecolor $12. ; markercolor = linecolor ; cards; 4 green 5 pink 6 yellow 8 red 10 blue ; run; proc format ; value cyl 4 = 'small' 5 = 'weird' 6 = 'large' 8 = 'extra large' 10 = 'ginormous' ; quit ; options orientation = landscape ; ods graphics / height = 8in width = 10in imagemap = on ; ods html5 path = "c:\deleteme" (URL=NONE) body = "deleteme.html" (title = "deleteme output") nogfootnote device = svg ; proc sgplot data = sashelp.cars dattrmap = cyl_attrs ; loess x = weight y = horsepower / group = cylinders attrid = cylinders ; xaxis grid ; yaxis grid ; where 4 le cylinders le 10 ; * format cylinders cyl. ; run ; run ; ods html5 close ; So, my grouping var here is a numeric. (In this dset that's an actual, ratio-level data element, but in my data it was at best ordinal. So it really didn't make sense to show the numeric values--I needed a format to make them interpretable.) I think it's counter-intuitive that the actual values I'm using to set the line & marker colors are completely different from the values I'm giving SAS to divide the points into groups. Given that the grouping var is a numeric, if I want that dattrmap to work I've got to change the value type to character & perfectly match all the text strings from my format--that cyl_attrs dstep has to be: data cyl_attrs2 ; retain id "cylinders" linepattern "solid" ; input @1 value $11. @13 linecolor $12. ; markercolor = linecolor ; cards; small green weird pink large yellow extra large red ginormous blue ; run; So now if I decide that 'ginormous' is unprofessional, and I want to change that format label to '2xl', I've got to remember to change the description both in the format and in the dattrmap dset. I think it'd be much cleaner if I could just set the value var to the exact type & domain as the grouping var.

RoyPardee · ‎02-23-2024

Heh--having just now run into this issue, a whole 6 years later, it seems to me this is undesirable behavior. It's counterintuitive, and keeps users from being able to mess with formatting w/out having to keep two different sets of values in sync. Anybody know the process for reporting a bug? I think this merits it.

RoyPardee · ‎12-17-2020

Yep, looks like this was fixed in one of those maintenance releases. Compare: to: Huzzah!

RoyPardee · ‎12-17-2020

Hmm--I'm on 1M4 right now, but do have access to a 1M6 install--will try there & see if I have different results.

RoyPardee · ‎12-16-2020

This code: options orientation = landscape ; ods graphics / height = 8in width = 10in ; %let out_folder = c:\deleteme ; ods html5 path = "&out_folder" (URL=NONE) body = "deleteme.html" (title = "deleteme output") style = magnify nogfootnote device = svg ; title1 "this is a title" ; proc sgplot data = sashelp.class ; hbox age / category = sex ; xaxis grid ; yaxis grid ; run ; run ; ods _all_ close ; Results in a title that's cut-off: This behavior is consistent across chrome and firefox. I've seen e.g., this post, but I really do want SVG output with integrated titles. Interestingly, if I take the contents of the <svg> element & save it off to an external .svg file, I can then insert that into a powerpoint slide & the title is not cut off. This is a bug, isn't it?

RoyPardee · ‎10-24-2019

Woah--never knew about yaxistable before--love that. I would also love the ability to set box width to indicate the number of obs in the category.

RoyPardee · ‎01-17-2018

Thanks much for the reply! Alas, that does not work for me--sas does not recognize the legenditem statement, no doubt b/c I'm not on the version you specified. Also--yikes--that's a lot of code. Maybe I should add it to the SAS Ballot as a requested enhancement...

RoyPardee · ‎01-17-2018

Hey All, I love the plot I get with the below code, except that I would like to have the different marker symbols appear on the lines in the legend--like this: (I get that if I specify a SERIES plot with the /MARKERS option.) Instead, I just get the colored lines w/no symbols. In the fake data I give below it's not a big deal, but my real graph has 12 different groups. I've played around with the KEYLEGEND statement, but have not figured out a way to add those symbols. Can it be done? I'm using 9.4 (TS1M4) on X64_7PRO. Many thanks! -Roy data gnu ; input @1 clinic $char5. @9 mo date9. @21 metric 3.0 ; format mo mmddyy10. ; datalines ; north 01jan2002 344 north 01feb2002 400 north 01mar2002 300 north 01apr2002 423 north 01may2002 600 north 01jun2002 603 south 01jan2002 356 south 01feb2002 412 south 01mar2002 500 south 01apr2002 489 south 01may2002 200 south 01jun2002 193 run ; options orientation = landscape ; * ods graphics / height = 8in width = 10in ; ods html path = "c:\deleteme" (URL=NONE) body = "deleteme.html" style = magnify nogfootnote ; proc sgplot data = gnu ; * series x = mo y = metric / group = clinic lineattrs=(pattern = solid) markers ; loess x = mo y = metric / group = clinic lineattrs=(pattern = solid) ; run ; ods html close ; The plot that gets me is:

RoyPardee · ‎05-18-2017

Hey All, Just discovered the HTML5 ODS output destination & created a lovely inline SVG with SGPLOT--yay! It looks great in chrome (Version 55.0.2883.87) but gets rendered as sad sad text in IE 11.0.9600.18638. I'm using SAS 9.4 (TS1M2) on the X64_7PRO platform, with the daisy style. Is this known/expected behavior? Is it fixed in an upcoming release maybe? Thanks! -Roy

RoyPardee · ‎10-24-2016

Thanks Shmuel! That solution is not super-appealing, unfortunately. In my real data I've got 87 rows I need to put out. I could actually run my dset through PROC TRANSPOSE I suppose, but I think I may just switch to PROC TABULATE for this. This code does what I want: proc tabulate data = have ; class site ; var row1-row2 ; tables (row1-row2)*sum="", site / misstext = '0'; run ; Output from that is: +-----------------------------------+--------------------------------------+ | | site | | +------------+------------+------------+ | | east | north | south | +-----------------+-----------------+------------+------------+------------+ |No. flu shots | | 23.00| 7.00| 5.00| +-----------------+-----------------+------------+------------+------------+ |No. broken arms | | 14.00| 1.00| 35.00| +-----------------+-----------------+------------+------------+------------+

RoyPardee · ‎10-24-2016

Hey All, I've got patient-level data grouped by site, like so: data have ; input @1 site $char5. @7 mrn $char6. @14 row1 2.0 @18 row2 2.0 ; label row1 = "No. flu shots" row2 = "No. broken arms" ; datalines ; east roy 2 12 east bob 21 . east mary . 2 north sara 0 1 north trevor 7 . south alice 1 0 south bart 1 0 south tim 0 31 south iggy 2 4 south gene 1 0 south don 0 0 run ; So--data in columns (vars) that I would like to report as rows. I can get pretty much a perfect transpose of the report I want with this: proc report data = have ; column site row1 row2 ; define site / group ; define row1 / 'flu shots' ; define row2 / 'arms' ; run ; That spits out: site flu shots arms east 23 14 north 7 1 south 5 35 But what I really want is more like: site east north south flu shots 23 7 5 arms 14 1 35 My impulse was to change site from a /group var to an /across var. But when I do that, I get just a single row, with row counts under each listed site, followed by individual columns for my row1 & row2 vars, with sums over all sites. I suspect that REPORT will easily do this for me, I just don't know how to make my wishes intelligible to it. Can one of you kind gurus enlighten me? Many thanks! -Roy

RoyPardee · ‎05-22-2014

Ah, that does the trick. Many thanks Sanjay! Here's what I came up with: data psa_percents ; input @1 clinic $char5. @7 provider $char15. @25 num_males75 3.0 @29 pct_overpsa 5.1 @35 dummy 1.0 ; datalines ; north Banderas, Antonio 80 15.2 1 north Jolie, Angelina 45 75.5 1 north Bale, Christian 130 30 1 west Claus, Santa 225 0 1 west Bunny, Easter 340 0 1 west Cruz, Ted 89 0 1 ; run ; proc template ; define statgraph psa_butterfly ; begingraph ; layout lattice _id='lattice' / columndatarange = data columngutter = 0 rowdatarange = data rowgutter = 10 /* columnweights = (0.62 0.38) */ columns = 2 ; layout overlay _id='overlay' / yaxisopts = (gridattrs=(color=CXE8E6E8 pattern=SOLID thickness=1 ) gridDisplay = on display =(LINE TICKVALUES TICKS)) xaxisopts = (label = '# PSAs 75+' reverse = true type = linear gridDisplay = on ) ; barchart _id='bar2' x=PROVIDER y=num_males75 / name='bar2' orient = horizontal ; endlayout; layout overlay _id='overlay2' / yaxisopts=(gridattrs=(color=CXE8E6E8 pattern=SOLID thickness=1 ) gridDisplay=on display=(LINE )) y2axisopts = (display = none) xaxisopts=(label="Percent of 75+ males in panel having PSA" type=linear gridDisplay = on ); barchart _id='bar' x=PROVIDER y= pct_overpsa / name='bar' orient = horizontal ; scatterplot _id = 'scat' x = dummy y = dummy / yaxis = y2 datatransparency = 1 ; endlayout; endlayout; endgraph; end; options orientation = landscape ; ods html path = "c:\temp" (URL=NONE) body = "deleteme.html" (title = "Axis Problems") ; proc print ; run ; proc sgrender data= psa_percents template = psa_butterfly ; by clinic ; run; ods _all_ close ;

RoyPardee · ‎05-22-2014

Thanks for the reply! Alas, the graphs are being shown separately, so I really want the x-axes to be scaled for each clinic (even the percentage one). I think the linearopts=(viewmax/min) option just lets me set a global min & max value that gets used accross all clinics.

RoyPardee · ‎05-21-2014

Hey All, I'm producing a 'butterfly graph' by-clinic, using code very like the below. My problem is that for clinics where all the providers have 0s for one of the stats I'm reporting, SAS is putting 0 smack dab in the middle of the x-axis, which looks really weird (please see attachment SGRender3.png, or run the below code to see what I'm talking about). The easiest way I can think of to fix this would be to specify min and max values for the x-axes. But I don't want to do that because the real data varies considerably by clinic, and there's no single max value I can use for every clinic that doesn't compress the provider-by-provider variation all to heck at some large proportion of the clinics. I asked this question on SAS-L and Joe Matise (god bless 'im) suggested coming and asking here. Is there some good way to keep SAS from letting that 0-value float into the center of my x-axes? Many thanks! -Roy data psa_percents ; input @1 clinic $char5. @7 provider $char15. @25 num_males75 3.0 @29 pct_overpsa 5.1 ; datalines ; north Banderas, Antonio 80 15.2 north Jolie, Angelina 45 75.5 north Bale, Christian 130 30 west Claus, Santa 225 0 west Bunny, Easter 340 0 west Cruz, Ted 89 0 run ; proc template ; define statgraph psa_butterfly ; begingraph ; layout lattice _id='lattice' / columndatarange = data columngutter = 0 rowdatarange = data rowgutter = 10 /* columnweights = (0.62 0.38) */ columns = 2 ; layout overlay _id='overlay' / yaxisopts = (gridattrs=(color=CXE8E6E8 pattern=SOLID thickness=1 ) gridDisplay = on display =(LINE TICKVALUES TICKS)) xaxisopts = (label = '# PSAs 75+' reverse = true type = linear gridDisplay = on ) ; barchart _id='bar2' x=PROVIDER y=num_males75 / name='bar2' orient = horizontal ; endlayout; layout overlay _id='overlay2' / yaxisopts=(gridattrs=(color=CXE8E6E8 pattern=SOLID thickness=1 ) gridDisplay=on display=(LINE )) xaxisopts=(label="Percent of 75+ males in panel having PSA" type=linear gridDisplay = on ); barchart _id='bar' x=PROVIDER y= pct_overpsa / name='bar' orient = horizontal ; endlayout; endlayout; endgraph; end; options orientation = landscape ; ods html path = "c:\temp" (URL=NONE) body = "deleteme.html" (title = "Axis Problems") ; proc sgrender data= psa_percents template = psa_butterfly ; by clinic ; run; ods _all_ close ;

Online Status	Offline
Date Last Visited	‎02-23-2024 08:09 PM

Re: sgplot scatter plot using attrmaps and formats

Re: sgplot scatter plot using attrmaps and formats

Re: titles cut off using html5/svg

Re: titles cut off using html5/svg

titles cut off using html5/svg

Re: Fun with SAS ODS Graphics: Does this box plot make my MWSUG 2019 p...

Re: SGPlot loess legend--can I get symbols?

SGPlot loess legend--can I get symbols?

HTML5 SVG output--no worky in IE11?

Re: PROC REPORT noob--need to kind of transpose my data

sgplot scatter plot using attrmaps and formats

Create a SAS Official Visual Studio Code Extension

Re: SAS macros in Git

Re: titles cut off using html5/svg

Re: title get cut off while creating a graph by sgpanel

Re: titles cut off using html5/svg

titles cut off using html5/svg

HTML5 SVG output--no worky in IE11?

SAS/ACCESS provider for Teradata should send/recieve variable labels f...

Re: sgplot scatter plot using attrmaps and formats

Re: sgplot scatter plot using attrmaps and formats

Re: titles cut off using html5/svg

Re: titles cut off using html5/svg

titles cut off using html5/svg

Re: Fun with SAS ODS Graphics: Does this box plot make my MWSUG 2019 p...

Re: SGPlot loess legend--can I get symbols?

SGPlot loess legend--can I get symbols?

HTML5 SVG output--no worky in IE11?

Re: PROC REPORT noob--need to kind of transpose my data

PROC REPORT noob--need to kind of transpose my data

Re: axis control with statgraph template

Re: axis control with statgraph template

axis control with statgraph template