Hi to all This is my table with 3 columns. id serial date 1 334 17.03.2015 2 334 18.03.2015 3 334 15.07.2015 4 334 19.09.2015 5 334 03.10.2015 1 335 10.02.2015 2 335 13.02.2015 3 335 18.03.2015 4 335 13.08.2015 5 335 05.10.2015   I want to get this: (new column [control] based on count of days between two date)   id serial date control 1 334 17.03.2015 OK 2 334 18.03.2015  3 334 15.07.2015 OK 4 334 25.07.2015 5 334 03.10.2015 OK 1 335 10.02.2015 OK 2 335 13.02.2015   3 335 18.02.2015  4 335 13.08.2015 OK 5 335 05.10.2015 OK   Rules and examples: 1. Serial with first (oldest date) or ID 1 is always control = 'OK' 2. Count days beetwen first and second date for serial 334 and if less then 14 days then live empty cell, if not then OK Continue until days count > 14, then start new count  example: (id 1 serial 334) 17.03.2015 - (id 2 serial 334) 18.03.2015 < 14  day so  control = ' ' 3. Count days beetwen first and third date for serial 334 example: ((id 1 serial 334) 17.03.2015 - (id 3 serial 334) 15.07.2015 > 14 days so   control = ' OK' 4. After reaching OK start to count fom that position.  example: ((id 3 serial 334) 15.07.2015 - (id 4 serial 334) 25.07.2015 < 14 days so   control = ' ' 5. Next: example:  ((id 3 serial 334) 15.07.2015 - (id 5 serial 334) 03.10.2015 > 14 days so   control = ' OK' Stop and go to the next serial Same for any other serial in table. Thank you in advance Bob       ... View more

Hi to all My data: data  have; input string $40.; datalines; 33_we#fff23BOB_TS=234623623624756_TOO BOB_TS=57335623566654756BO 60pBOB_TS=343467364563423TOp_# WWBOB_TS=2346752665634cc90 9789BOB_TS=9872363566756Alll ldldBOB_TS=2340235674025634efrwe ; run; I want to extract first 10 digits after string 'BOB_TS=' (example:  have 60pBOB_TS=343467563423TOp_#  want 3434675634 ) Ca somebody help me with this? Thank you in advance Bob ... View more

Hi to all Can someone help with this? My problem is in proc gplot  and plot column "operations" As you can see from example below every time I must manually add value for plot (a*week=2 b*week=3 ....) For example if value a is missing, in some period of time, I must rewrite it with existing otherwise SAS will show an error.  Is there any solution to avoid manual intervention for every output? Thank you in advance Bob data test; length operation $1; format percent percentn10.2; input week operation percent sum_week; datalines; 18 a 0.21 5842 18 b 0.34 5842 18 c 0.13 5842 18 d 0.15 5842 18 e 0.02 5842 18 f 0.11 5842 19 a 0.11 8501 19 b 0.14 8501 19 c 0.23 8501 19 d 0.25 8501 19 e 0.32 8501 19 f 0.21 8501 20 a 0.21 7007 20 b 0.34 7007 20 c 0.13 7007 20 d 0.15 7007 20 e 0.02 7007 20 f 0.11 7007 ; run; proc sort data=test (drop=sum_week) out=test1; by week; run; proc transpose data=test1 out=test1 (drop=_name_); by week; id operation; run; proc sql; create table test2 as select unique week, sum_week from test; create table test_final as select unique test1.*, test2.sum_week from test1 left join test2 on test1.week=test2.week; quit; run; axis1 label=('Lines') order=(0 to .40 by .05) minor=none offset=(0,0); axis2 label=none minor=none offset=(15,15); axis3 label=('Bars') order=(0 to 9000 by 1000) minor=none offset=(0,0); symbol1 value=none interpol=needle width=25 color=cx4573a7 repeat=1; symbol2 value=none interpol=join repeat=1 color=blue; symbol3 value=none interpol=join repeat=1 color=cx00ff00; symbol4 value=none interpol=join repeat=1 color=pink; symbol5 value=none interpol=join repeat=1 color=red; symbol6 value=none interpol=join repeat=1 color=purple; symbol7 value=none interpol=join repeat=1 color=black; legend1 position=(bottom center) down=1 label=none; proc gplot data=test_final; format a percent7.0; plot a*week=2 b*week=3 c*week=4 d*week=5 e*week=6 f*week=7 / overlay vaxis=axis1 haxis=axis2 legend=l1gend1; plot2 sum_week*week=1 / vaxis=axis3 haxis=axis2; run; ... View more

Hi can somebody help mi with this. My question is how to put two dimension  on horizontal axis wk and year to get proper order from wk 42/2013 to 1/2014. This script was fine until first week of 2014. I tried to make new variable with compress (year|| "*"||wk)  and then to split axis by * but there was error : All variables must be the same type for overlay request. Thank you in advance data pass_fail;                                                                                                                            input wk actual pass fail year;                                                                                                   datalines; 1 423      423 0 2014 42 6929 6915 14 2013 43 9344 9334 10 2013 44 7329 7322 7 2013 45 8263 8248 15 2013 46 8266 8252 14 2013 47 7525 7520 5 2013 48 10299 10276 23 2013 49 11059 11056 3 2013 50 10444 10443 1 2013 51 11776 11776 0 2013 52 5015 5014 1 2013 ; run; data plan; set pass_fail; plan = 8400;  format ratio percent7.; ratio = pass / (pass + fail ) ; week=wk-.13; run; goptions reset=ALL gsfmode=replace htext=30pt cback=white border htitle=32pt htext=16pt ftext="Arial/bold"; axis1 label=(a=90 h=25pt 'Actual /Plan') order=(0 to 12000 by 2000) minor=none offset=(0,0); axis2 label=(a=90 h=25pt'Ratio') order=(.82 to 1.00 by .01) minor=none offset=(0,0); axis3 label=( h=25pt'Week') minor=none offset=(0,10) value=(t=1 ' '); symbol1 value=dot cv=purple height=2 interpol=join width=5 color=purple ; symbol2 value=none interpol=needle width=50 color=red; symbol3 value=none interpol=needle width=50 color=blue; footnote h = 25pt c = black 'Actual ' c=Red f=marker 'U' h = 25pt  c=black f='Arial/bold' '  Plan  ' c=blue  f=marker 'U' h = 25pt  c=black f='Arial/bold' '  Ratio  ' c=purple   '____'; proc gplot data=plan; plot actual*wk=2 plan*week=3 / overlay vzero vaxis=axis1 haxis=axis3 autovref cvref=graydd; plot2 ratio*wk=1 /  vaxis=axis2 ; run; quit; ... View more

Hi Is it possible to run script like one below in SAS (9.1) without adjusting to SAS SQL language? SELECT TO_CHAR (reg_date, 'YYYYIW') yyyyiw,          model_no,          material_no,          MAX (wtd) wtd,          MAX (mtd) mtd,          MAX (ytd) ytd     FROM (SELECT sr.reg_date,                  sr.part_no model_no,                  ip.type_designation material_no,                  COUNT (1) OVER (PARTITION BY TO_CHAR (sr.reg_date, 'YYYYIW'), sr.part_no)                     wtd,                  COUNT (1) OVER (PARTITION BY TO_CHAR (sr.reg_date, 'YYYYMM'), sr.part_no)                     mtd,                  COUNT (1) OVER (PARTITION BY TO_CHAR (sr.reg_date, 'YYYY'), sr.part_no)                     ytd             FROM amb.service sr                  INNER JOIN                  amb.part ip                     ON     ip.contract = sr.part_contract                        AND ip.part_no = sr.part_no            WHERE     sr.part_contract = '12506'                  AND sr.agreement_id <> 'PACEF12506'                  AND sr.state <> 'New')    WHERE TO_CHAR (reg_date, 'YYYYIW') = TO_CHAR (SYSDATE-2, 'YYYYIW') GROUP BY TO_CHAR (reg_date, 'YYYYIW'), model_no, material_no; Thanks Bob ... View more