I am note sure how to post this, but here goes. I am using airline navigation data and I need to organize the data into a time series format for analysis. That is, for each monthly date I have a set of variables and associated data. The data is organized into flight zones. They are Vancouver, Edmonton, Winnipeg, Toronto, Montreal, Moncton, Gander, and Gander_OCA. For each region there are two different types of flights; IFR and VFR. Now making time series based on region and type is easy...the problem I have is adding markets. We have up to 13 markets, but not all regions will experience flights from the 13 markets, but I need to future proof in case such a flight comes. Each region will have hours in the region and number of flights. I can write macro loop for the region and type, but how do I make variables for each of the market types, and if no data exists, still create the market but populate it with zeros. Here is the code I use to make the time series without any markets:   ods listing close; ods html close; goptions device=actximg; %macro fir_all (var_list=); %local n fir_var; %do n=1 %to %sysfunc(countw(&var_list)); %let fir_var=%scan(&var_list,&n); /* Get IFR Flights */ PROC SQL; CREATE TABLE EGTASK.&fir_var._ARRANGE_IFR AS SELECT t1.Date, t1.FIR_NUF LABEL="Non-Unique &fir_var IFR Flights" AS &fir_var._FIR_IFR_NUF, t1.FIR_HOURS LABEL="&fir_var FIR IFR Hours" AS &fir_var._FIR_IFR_HOURS FROM EGTASK.TOTALS_RULE_FIR t1 WHERE t1.FIR_NAME = "&fir_var" AND t1.Flight_Rule_Code = 'IFR'; run; /* Get VFR Flights */ PROC SQL; CREATE TABLE EGTASK.&fir_var._ARRANGE_VFR AS SELECT t1.Date, t1.FIR_NUF LABEL="Non-Unique &fir_var VFR Flights" AS &fir_var._FIR_VFR_NUF, t1.FIR_HOURS LABEL="&fir_var FIR VFR Hours" AS &fir_var._FIR_VFR_HOURS FROM EGTASK.TOTALS_RULE_FIR t1 WHERE t1.FIR_NAME = "&fir_var" AND t1.Flight_Rule_Code = 'VFR'; run; /* Join IFR and VFR Flights into one dataset */ PROC SQL; CREATE TABLE EGTASK.&fir_var._FIR AS SELECT t1.DATE, t1.&fir_var._FIR_IFR_NUF, t1.&fir_var._FIR_IFR_HOURS, t2.&fir_var._FIR_VFR_NUF, t2.&fir_var._FIR_VFR_HOURS FROM EGTASK.&fir_var._ARRANGE_IFR t1 LEFT JOIN EGTASK.&fir_var._ARRANGE_VFR t2 ON (t1.DATE = t2.DATE); proc datasets nodetails nolist nowarn lib=egtask; delete &fir_var._ARRANGE_IFR &fir_var._ARRANGE_VFR &fir_var._ARRANGE_UNK; run; %end; %mend; %fir_all (var_list=VANCOUVER EDMONTON WINNIPEG TORONTO MONTREAL MONCTON GANDER GANDER_OCA);   I am now stuck with integrating the market data. There are 13 markets called "Market 1", "Market 2", etc...but again, not all regions will get all flights, but they MAY get them all in the future. So I need to be able to merge in this data, even if a region has none. If there is none, then create the variable (e.g., Edmonton-IFR-Market1) and set it to zero. In the end I need a dataset like:   Jan2000 Edmonton-IFR-Flights-Market1 to Market13 Edmonton-IFR-Hours-Market1 to Market13 Edmonton-VFR-Flights-Market1 to Market13 Edmonton-VFR-Hours-Market1 to Market13    Each region will have 13 IFR data variables and 13 VFR data variables for both flights and hours in the time series. All regions need to have all variables populated. If they do not have data, then create the variable and set observations to zero (or missing, as I can set missing to 0 later).   Not sure how difficult this is...but I cannot figure it out (checking if variable exists or not).  Thanks and Happy New Year. ... View more

Hi all, I have an issue with adjusting the look of my data. I am using EG but can also program reasonably well.   I have data ordered as:   Customer_ID     Invoice_num     Customer_name     Chrg     Amt 1121                        1                    Bank1                    Fee1       1                                                                                     Fee2        2                                                                                     Fee3        3 1122                        2                    Bank 2                   Fee4       12                                                                                     Fee7       15   And so on. There are up to 9 Fee Chrgs. (fee is a description of the fee type)   I need to flatten this so I get   Customer_ID     Invoice_num    Customer_name    Chrg1 Chrg2 Chrg3 Chrg4 (up to 9)  Amt1 Amt2 Amt3 (up to 9) and a total 1121                          1                    Bank1                Fee1  Fee2   Fee3  Fee4                     1       0       5      1122                          6                    Bank2                Fee1  Fee2   Fee3  Fee4                     1       0       5        All Fee description types (9) must be listed horizontally, if a client does not have that fee type on their invoice, it would be 0   I am a bit stuck with this.  Thanks ... View more

I have some data that is organized into seasons. I would like to test for significant differences between the seasons (spring summer fall winter). I can generate the means fine:   ods graphics on; proc sort data=tafsas.nav_data_all_ts; by season_name; proc means data=tafsas.nav_data_all_ts noprint; var ovr_ecu; by season_name; where date>='01Jan2010'd; output out=meanstest mean=; run; ods graphics off; So I am not sure how to proceed from here.  Any advice is appreciated.     Thx. ... View more

Hi all,   Not sure this is the best sub-forum, so mods please move if you feel necessary.   We have some SAS projects that are run via Windows Scheduler. They run for a bit, but then run for a few seconds and do nothing. These jobs typically take about 30 mins. to run. But it seems that after about a week of successful runs, they initiate, say they run, and finish, all in about 9 seconds.   Has anybody heard of this happening and any resolutions? This is very weird.   Perhaps a Windows Scheduler alternative?   thanks ... View more