If I have 3 macro codes for 3 different types of data sets:   %Converter1(File, Start_Date) %Converter2(File, Start_Date) %Converter3(File, Start_Date)   Then I would want my master macro code to be something like:   %Formatter(File, Start_Date, Converter_Type).   So that if I were to write:   %Formatter(//FILELOCATION.xlsx, 2AUG2016, Converter1)   It would run %Converter1.   So then my %Formatter Code would look something like   %Formatter(File, Start_Date, Converter_Type) / store source;   if Converter_Type = &Converter_Type, then %Converter1(&File, &Start_Date) if Converter_Type = &Converter_Type , then %Converter2(&File, &Start_Date) if Converter_Type = &Converter_Type , then %Converter3(&File, &Start_Date)   Does that make sense?        ... View more

I am unfortunately forced to work in enterprise guide but I only use the coding feature. To be honest, I don't even know how to do the click features. So yes, I am comfortable writing code.    The description is simple....    There are X amounts of initial formats for data that I am processing. Currently, I have a different code for each of the types of data sets. Instead of having X amounts codes, I would rather have one. That way if there are 5 different types of programs/codes that are filtering and formatting the data, I can only call ONE program and indicate which program it should run based on a macro call.    So lets say I write a macro program which I am very comfortable with and it is %format_converter.   And I have 3 different Data set formats    Data Format 1: Text Format Pacific Time Zone  No Customer IDs Visitors   Data Format 2: Excel Format Eastern Time Zone Customer IDS Visitors Orders    Data Format 3: Text Format Eastern Time Zone No Customer IDs Visitors Orders    The final format of all of these files should be:   Final Format: CSV Format Eastern Time Zone Visitors Per Minute Orders Per Minute     %macro format_converter( File, Start_Date, End_Date, Data_Format Type) / store source   There will be 3 choices for Data_Format Type and it will dictate which macro is called to format the data set.   Data set 1: Needs to be read in with infile statement Needs to be converted to Eastern Time No CIDs and no orders so very simple, just sum visits by minute. Orders column will be 0 FINISHED Data set 2: Needs to be read in with proc import Already in Eastern Time so no conversion necessary. CID’s on the visit and order data track which customers bought what. If visit comes in a 3:05pm and the order comes in at 3:56 pm, the order will be sourced back to 3:05 Visits and SOURCED orders compiled by minute. FINISHED Data set 3: Needs to be read in with infile statement Already in Eastern Time so no conversion necessary There are orders and visits but no CIDs Visits and UNSOURCED orders are compiled by minute Finished     So as you can see the data is similar for a lot of the formats but since they differ slightly they need different formatting and prepping for analysis.   Currently, I have ALL of these codes created and am using them daily in order to manage the files that are coming in. Instead of having all these codes that are doing ALMOST the same thing except for minor differences, I would like to create ONE code, %format, and then essentially say: If this format then follow these instructions If that format then follow those instructions ….     ... View more

Yes, I know this will require a macro....   The files are all the same format so they are structured perfectly. They were actually originally in one large files but then split out based on the day.    The reason I did not leave them all in one file is because I will be using a rolling-window to look at customers.      ... View more

Rest assured... all the original times are being kept.   I'm running vector autoregressions with literally thousands of exogenous variables that at times occur simulatenously or very close to each other in time....   For my purposes, if I have 10 exogenous effects all occuring within 3 minutes of each other they are all occuring at the same time.    For example, if I have ten exogenous variables that occur at 3:00, then instead of 10 variables I have one variable....which is defined as 10 exogenous variables at 3:00. At this point of my analysis, I am not worried about the individual effect of all of those variables...just the combined effect. I worry about the individual effects later in my analysis.    So I know how to combine the variables when tehy occur at the same time...but I need to know how to combine them when they are one, two, or three minutes away from each other.    Garrett ... View more

Hi Ballardw,    Thanks for your response. I know exaclty what I want to do but I don't know the name of the concept.    Here is an example....    I have three times...  1: 3:00 2: 3:02 3: 3:03   For the purpose of my analysis, these three times should be the same if they are less than or equal to 3 minutes apart from one another and they should be the time of the earliest observation.    The time between 3:02 and 3:00 is 2 minutes. The time between 3:03 and 3:02 is 1 minute. The time between 3:03 and 3:00 is 3 minutes.    If I recalculate with code such as:   if Time-lag(Time) <=3 then Time = lag(Time);   Then my recalculated times would be: 1. 3:00 2. 3:00 3. 3:02   This is because the condition in observation 3 is based on the original value of 3:02 and not the re-calculated value of 3:00.    2) If 3:02-3:00 <=3 then Time = 3:00. CONDITION IS MET AND TIME IS RE-CALCULATED TO 3:00. 3) If 3:03-3:02 <=3 then Time = 3:02. CONDITION IS MET AND TIME IS RE-CALCULATED TO 3:02.   But I want my times to be:   1. 3:00 2. 3:00 3. 3:00   Does that make sense?  ... View more

The data is read into a SAS dataset. I just used excel for this purpose.    Yes, they do need to be in a macro variable.    The data is re-run weekly so the transactions will change if there are no orders or revenue or others... (there are over 15 different transactions")   The data is originally read in a text format and I use a scan function to read the values that exist in the "Transacation Type" field and then use a do loop to create columns for each transaction. So if I specify the Transaction Types without doing a scan, I run the risk of losing transactions that are different from one week to the next.    They are all numeric.  ... View more