If you're allowed to modify the original datasets, something like this might work ...   data a;  length str $100.;  set a;  run;   data b;  length str $100.;  set b;  run;   data c;  length str $100.;  set c;  run;   proc sql ; create table want as  select * from a  outer union corr select * from b  outer union corr  select * from c  outer union corr  ;quit;  ... View more

Are you allowed to edit the Excel file? If so, try creating an extra column, and add any "explanatory text" (such as "avec un mini de 50...") to that extra column. Once the numeric data column has only numeric data in it, it should be much easier for SAS to import.   ... View more

If the 4 graphs can share the same axes, I recommend Proc SGpanel.   If you want the 4 graphs to be independent, I recommend Proc GReplay, with 4 Proc GPlot's. Here's an example of Greplay: http://robslink.com/SAS/democd67/anscombe_quartet.htm https://robslink.com/SAS/democd67/anscombe_quartet.sas ... View more

I suspect that one of the two maps doesn't have enough points along the border to follow the desired projection's curve. To try to fix it, you could either use a less-curvy projection (such as cylindri), or you could leave more points in the map (by eliminating the where clause's "density<=something").   The longer explanation:    The reason it doesn't have enough points is (probably) that when the maps were processed by SAS (before  being shipped to the customers), Proc Greduce was probably run against the unprojected coordinates ... and the coordinates along that part of the US/Canada border pretty much have the same latitude, and therefore Proc Greduce thinks it doesn't need to keep very many data points along that border. But once you start using a projection, and those borders are shown as 'curves' then you need more coordinates to follow the curve.   One of the final projects I inherited and worked on before retiring was the processing of the maps. I changed the algorithm to project the map first, before running Proc Greduce, to try to reduce/eliminate this problem. I believe that the new maps created using my algorithm are the ones currently available for download (and are likely newer than the ones you have in your mapsgfk library). You might try downloading them, and see if it gives you better resolution along that border: https://support.sas.com/en/knowledge-base/maps-geocoding/maps-online.html   ... View more

I'm not 100% sure, but you'll probably have to modify the template something like this... https://support.sas.com/kb/34/186.html   ... View more

In addition to what @ballardw pointed out, I would also suggest using the new maps from the mapsgfk library rather than the (possibly outdated) maps from the old maps library. One caveat there is that SAS only ships 1 Canada map in mapsgfk, therefore you'll have to derive the province map from it (using proc gremove) - that can be a little confusing if you haven't done it before, therefore I'm including some code below:   data map_us; set mapsgfk.us_states (where=(statecode not in ('AK' 'HI') and density<=4)); my_id=statecode; country='US'; run; data map_canada; set mapsgfk.canada (where=(id1 not in ('CA-60' 'CA-61' 'CA-62') and density<=1)); run; proc gremove data=map_canada out=map_canada; by id1; id id; run; data map_canada; set map_canada; my_id=id1; country='CA'; run; data mapall; length my_id $8; set map_us map_canada; run; proc sql noprint; create table mapall_attr as select unique country, my_id, 1 as fakedata from mapall; quit; run; proc gproject data=mapall out=mapall_proj latlong eastlong degrees project=gnomon; id country my_id; run; pattern1 v=solid color=white; proc gmap map=mapall_proj data=mapall_attr all; id my_id; choro fakedata / levels=1 nolegend coutline=gray55; run; ... View more

Here's a slightly modified version of the code, that uses lat/long values until the end, and then projects them to projected x/y values at the very end. The original SGF paper example used maps.states and deals with converting between degrees and radians, whereas I use mapsgfk.us_states and keep the coordinates in degrees until I project them.   %macro make_anno(annodata,inputdata, rwidth,rdepth,lcolor,fcolor,solid,center,tipvar); data &annodata; length function $8 color $ 12 style $20 ; retain xsys ysys '2' hsys '3' when 'a' size .7; /* default grid block size variables, if not passed in (in degrees) */ width=1; depth=1; %if (&inputdata ^= ) %then %do; set &inputdata; %end; %if (&solid = 'Y') %then %do; style='solid'; %end; %else %do; style='empty'; %end; line=1; /* optional arguments */ %if (&rwidth ^= ) %then %do; width=&rwidth; %end; %if (&rdepth ^= ) %then %do; depth=&rdepth; %end; /* optional tool tips */ %if (&tipvar ^= ) %then %do; html= 'title=' || quote("&tipvar" ||'=' || trim(left(&tipvar)) ); %end; start_long =long; start_lat= lat; /*** create the rectangle ***/ /* coordinates of the rectangle */ function='POLY'; color=&fcolor; long=start_long-width; lat=start_lat+depth; output; /* start position */ function='POLYCONT'; long=start_long+width; lat=start_lat+depth; output; long=start_long+width; lat=start_lat-depth; output; long=start_long-width; lat=start_lat-depth; output; run; %mend; /* get min & max values, to use for scaling colors */ %macro get_minmax(inds,varname); %global minval; %global maxval; data tempmm; set &inds END=lastob; retain maxval -999999999; retain minval 999999999; if &varname <= minval then minval=&varname; else if &varname >= maxval then maxval=&varname; if (lastob = 1) then do; output; call symput("minval",minval); call symput("maxval",maxval); end; run; %mend; %macro draw_map(mapdata,respdata, annodata, id, choro, name, dev, mapcolor, htmlvar); GOPTIONS reset=all; GOPTIONS CBACK=white; GOPTIONS DEVICE=&dev xpixels=700 ypixels=500; ODS LISTING CLOSE; %if ((&mapcolor ne ) and (&mapcolor = "N")) %then %do; pattern v=me c=CX000000 r=100; %end; %else %do; pattern v=s c=CXE9E8DC r=100; %end; %if (&htmlvar ^= ) %then %do; %let htmlv=html=&htmlvar; %end; %else %do; %let htmlv=; %end; goptions border; proc gmap map=&mapdata anno=&annodata data=&respdata all; id &id; choro &choro /nolegend &htmlv; run; quit; %mend; data my_data; infile datalines dsd truncover; input orientation:$10. Target:$15. k:32. station:$8. lat:8.1 long:8.1 xpos1:32. ypos1:32. ksum:32. ksumk:32.; format lat 8.1 long 8.1; datalines; 45/35,360/-30/100,20,USL000,34.5,-110.7,360,-30,3,23 45/35,390/-30/100,18,USL000,34.5,-110.4,390,-30,3,21 45/35,420/-60/100,20,USL000,34.2,-110.1,420,-60,4,24 45/35,450/-60/100,20,USL000,34.2,-109.7,450,-60,4,24 45/35,450/-30/100,19,USL000,34.5,-109.7,450,-30,3,22 45/35,480/-90/100,19,USL000,34.0,-109.4,480,-90,4,23 45/35,480/-60/100,19,USL000,34.2,-109.4,480,-60,4,23 45/35,480/-30/100,19,USL000,34.5,-109.4,480,-30,3,22 45/35,510/-90/100,19,USL000,34.0,-109.1,510,-90,4,23 45/35,540/-120/100,17,USL000,33.7,-108.8,540,-120,4,21 ;;;; run; %let oildata=my_data; %let predvar=ksumk; %get_minmax(&oildata,&predvar); /* set the color for each data point */ data &oildata; set &oildata; length hx $2; /*make negative numbers green to black */ /*the more negative the number, the greener */ if (&predvar < 0) then do; cnum=(&predvar * (255/&minval)); /*scale colors */ /*Make sure we stay within the threshold values*/ if (cnum < 0) then cnum=0; else if (cnum > 255) then cnum=255; hx=put(cnum,hex2.); /* Transparency for 9.3 and up*/ color="A00"||trim(left(hx))||"0099"; /* Pre-9.3 cannot use transparency **** color="CX00"||trim(left(hx))||"00"; ***/ end; /* Make Positive values black to Red */ /* Larger numbers are more red */ else do; cnum=(&predvar * (255/ &maxval)); /*Make sure we stay within the threshold values*/ if (cnum <= 0) then cnum=0; else if (cnum > 255) then cnum=255; hx=put(cnum,hex2.); /* Alpha-transparent colors */ color="A"||trim(left(hx))||"000099"; end; run; %make_anno(annogrid,&oildata,.15,.15,color,color,'Y','N',&predvar); run; data mystates; set mapsgfk.us_states (where=(statecode in ('TX' 'OK' 'KS' 'CO' 'NM' 'AZ' 'UT' 'NV' 'CA'))); run; /* project the map and the annotate data using the same parameters */ proc gproject data=mystates out=mystates latlong eastlong degrees project=hammer parmout=parms; id id; run; proc gproject data=annogrid out=annogrid latlong eastlong degrees parmin=parms parmentry=mystates; id; run; %draw_map(mystates, mystates, annogrid,state,state,oil,png,"N"); run; ... View more

goptions xpixels=300 ypixels=100; proc gkpi mode=basic; hslider actual=.32 bounds=(0 .50 .70 1.00) / format="percent7.0" noavalue bfont=(h=12pt) colors=(cxdd5757 cxffcc32 cx15b57b); run; quit;   ... View more

Do you have a screen-capture (or even a hand-drawn image) of an axis, the way you're wanting it?   ... View more

Here's an example using SAS/Graph proc gplot ...   data ft2; input ftdy y; datalines; -48 -10 1 2 15 5 29 7 57 12 85 16 ; run; axis1 label=none minor=none offset=(0,0) order=(-20 to 20 by 5); axis2 label=none order=(-48,1,15,29,57,85) offset=(5,5) value=( t=1 'Screening' t=2 'Baseline' t=3 'Visit 4 (Day 15)' t=4 'Visit 5 (Day 29)' t=5 'Visit 6 (Day 57)' t=6 'Visit 7 (Day 85)' ); title1 "Drugs, Inc."; proc gplot data=ft2; plot y*ftdy / vaxis=axis1 haxis=axis2; run;     ... View more