Hi I have some data that looks like: order, businessname 1234,Acme 432,Bizarro 55689,Ralph's party town In the title I mentioned that the variables have different lengths and column starts.  By this I mean that the order numbers have different numbers of characters for each order and that the businessname observations start at different column positions.  I started my infile with a data step: data order; infile 'location.csv' dlm=',' firstobs=2; input order businessname $; run; SAS infiles the first 8 characters of businessname but I am demanding and want the whole enchilada.  If I knew the start column of businessname I would try to grab a field corresponding to businessname, i.e. something like $ 7-26 (Ralph's party town).  The problem is that the Acme observation starts at column 6, Ralph's at column 7 and Bizarro at 5.  How can I code an input statement for the businessname observations to flexibly accomdate this yet also grab all the characters in each observation? thank you for your help ... View more

Howdy I have a file of employment data corresponding to 9 counties that I wish to aggregate across counties by NAICS.  Within the file are employment observations for many different NAICS for each county.  I wish to aggregate the employment obs. over all the counties by NAICS.  I can do this with proc means or proc summary.  There is a problem though.  Some of the counties have observations that are missing.  If I use proc means to aggregate the data, it treats the missing values as zeros.  I would like to output an aggregated dataset for only those NAICS that have employment values for all nine counties.  If a single county has a missing value I do not want to aggregate the employment observations.  My thought is to subset the original data so that it contains data on NAICS with observed values for all nine counties.  I know how to drop observations if a observation is missing but I do not know how to drop an entire class (NAICS) if one or more county observations in that class are missing. Is there some exceptional SAS-pert that can help me? ... View more

howdy i have some code for two histograms and each displays in a separate graph, I'd like to overlay them and cannto figure out how to do it.  Here is my code for the two histograms: title '1995:Q1 Age Distribution'; proc gchart data=manu_agedist(where=(time="1995:1" & agegrpfm^="14-99")); vbar agegrpfm / discrete type=percent sumvar=emp sum; run; title '2000:Q1 Manufacturing Employment Age Distribution'; proc gchart data=manu_agedist(where=(time="2000:1" & agegrpfm^="14-99")); vbar agegrpfm / discrete type=percent sumvar=emp sum; run; Do i need to create two data subsets, one for time=1995:1 and then again for time=2000:1, merge them into another dataset with two series and the plot each series?  Is there a way I can use proc gchart to graph each series and overlay them from one dataset? How do I insert series labels in a chart with multiple series? thx much ... View more

Hi I have data like the following: quarter area naics agegrp hires 2001 1 1 11 1 604 2001 1year 1 11 2 416 2001 1 89 11 1 55 2001 1 89 11 2 86 2001 1 1 51 1 203 2001 1 1 51 2 143 2001 1 89 51 1 12 2001 1 89 51 2 8 2001 2 1 11 1 300 2001 2 1 11 2 256 2001 2 89 11 1 45 2001 2 89 11 2 67 2001 2 1 51 1 56 2001 2 1 51 2 76 2001 2 89 51 1 12 2001 2 89 51 2 7 2001 3 1 11 1 412 2001 3 1 11 2 365 2001 3 89 11 1 42 2001 3 89 11 2 56 2001 3 1 51 1 532 2001 3 1 51 2 432 2001 3 89 51 1 34 2001 3 89 51 2 12 I am trying to create something like: year quarter naics hiresN 2001 1 11 1161 2001 1 51 366 2001 2 11 668 2001 2 51 151 2001 3 11 875 2001 3 51 1010 where hiresN is the sum of hires in the first data set over counties and age groups How could I code this? ... View more

Hi, I have a variable (agegrp) that is an age interval identifier.  Values of agegrp would look something like 14-18, 19-21, etc.  My second variable is a count of employment (emp) that corresponds to agegrp identifiers.  I am trying to create a histogram that is an age distribution of employment.  I'd like to graph an age distribution with each series the share of total employment x agegrp.  On the y-axis would be percent of total employment and the agegrp identifiers on the x-axis.  The code I have so far is: proc gchart data=manu_agedist(where=(time="1995:1")); vbar agegrp / dsicrete type=percent                     sumvar=emp sum; run; This code yields counts of employment on the y-axis and agegrps on the x-axis.  Does anyone have advice on how to fix my code so I can graph an age distribution with each series the share of total employment x agegrp? ... View more

Hi I have two variables I'd like to concatenate (year, quarter) but using "||" to concatenate the variables leaves a whole bunch of spaces in the created variable.  I am looking to create a year-quarter identifier without the spaces in the new variable.  I would like to create a year-quarter variable like: 00:1 but I am having some trouble. Here is what I was trying to do to get that variable: data blah; set blah2; yr=substr(year,3,2); yearquarter=yr||":"||quarter; run; In the log there is a line that says: "numeric values have been converted to character values at the places given by: 168:11 169:22. The yr variable has missing observations and the year quarter variable has what looks to be a tab btwn the : and the quarter identifier.  Any help would be appreciated. ... View more

Hi I am trying to create a new variable that contains observations that are summed observations from another variable conditional on if the observations fall into a class. Example: Class: NAICS Old variable: employment New variable: employment_subset logic to create new variable: employment_subset = sum of employment where NAICS in(21,22,23,..) I have used proc means to do something similar, e.g. collapse observations by class.  Here I want to collapse observations over a specified range categories within a class.  Not sure how I would do this either with proc means or in the data step.  Can I do this with proc means?  If so, how do I condition the collapse over a range of categories within the class? thanks  ... View more

I have a monthly time series and want to make quarterly observations. I started by assigning observations to quarters and thought to somehow use proc means to generate the average value of each trio of year-quarter values. How do I use proc means to create mean values where of all like year and quarter values? data structure: year month quarter value 1992 1 2 x 1992 2 2 y 1992 3 2 z 1992 4 3 a 1992 5 3 b 1992 6 3 c Goal: use proc means to calc: (x + y z) / 3 , (a + b + c) / 3 ... View more

I have observations for one variable and two additional variables representing two separate classes. I would like to create a variable D, representing a ratio that relates an observation for variable var A in class B = 1 to observation for var A in class B = 0 given that A and B are in class C. The data set might look something like: varA varB varC 1 1 1 2 1 1 3 0 1 4 1 2 5 1 2 6 0 2 7 1 3 8 1 3 9 0 3 varD: (1/3) (2/3) (4/6) (5/6) (7/9) (8/9) thanks much for any help ... View more