About devsas

devsas · ‎10-17-2017

319 array diff{*} diff_all_mean-diff_all; ERROR: Missing numeric suffix on a numbered variable list (diff_all_mean-diff_all). WARNING: Defining an array with zero elements. 320 do i=1 to dim(all);

devsas · ‎10-17-2017

Yes, that was a mistake. But I defined it now and still get errors- data fish_f; set fish2_; array all{*} hg n elv sa z da rf fr; array all_mean{*} hg_mean n_mean elv_mean sa_mean z_mean da_mean rf_mean fr_mean ; array diff{*} diff_all_mean-diff_all; do i=1 to dim(all); diff[i]= all_mean{i} - all{i}; end; run;

devsas · ‎10-17-2017

Ok, so it seems I have nailed the first two steps using proc means and data step merge, but the last step using arrays im still not getting it. Here is my complete code- Proc means data = fish missing nway noprint ; class lt/missing; class dam/missing; var hg n elv sa z da rf fr ; output out = statz2 mean = median = /autoname; run; data fish2_ (keep = name hg n elv sa z da rf fr hg_mean n_mean elv_mean sa_mean z_mean da_mean rf_mean fr_mean hg_median n_median elv_median sa_median z_median da_median rf_median fr_median) ; merge fish (in = x) statz2 (in = y); by lt dam; if x = 1; run; *using arrays to calculate mean difference only; data fish_f; set fish2_; array all{*} hg n elv sa z da rf fr; array all_mean{*} hg_mean n_mean elv_mean sa_mean z_mean da_mean rf_mean fr_mean ; do i=1 to dim(all); diff[i]= all_mean{i} - all{i}; end; run; Please note that I only want to use data step and arrays to solve this problem

devsas · ‎10-17-2017

Yes, thanks you are right. I used the nway option and got the same results. Im still trying to understand though what it means. Doing some reading, but still not clear.

devsas · ‎10-17-2017

Ok, I tried doing it with data step and arrays, but got stuck again. Here is my data step merge code, not sure how can I integrate array statement here to calculate mean, median differences and percent differences? data fish2_ (keep = name hg n elv sa z da rf fr hg_mean n_mean elv_mean sa_mean z_mean da_mean rf_mean fr_mean hg_median n_median elv_median sa_median z_median da_median rf_median fr_median) ; merge fish (in = x) statz2 (in = y); by lt dam; if x = 1; run;

devsas · ‎10-17-2017

Thanks PGSTATS, your solution works too, its just that for this exercise I want to do it using proc means and arrays only to understand these better.

devsas · ‎10-17-2017

Thanks so much! I see that proc summary gets pretty much the same results as my code using proc means do, except that _type_ = 3 is the only one in your results. Can you please explain the significance of that and how can I get the same result using proc means? I believe I will do the merge using the data step and use arrays for repetitive statements.

devsas · ‎10-16-2017

Ok, so here is the situation- I have a dataset called Fish which has data onfish in 120 lakes, detailed in 120 rows and 15 columns including lake name. The scope of my project is to compare each lake to other similar lakes to see whether there are differences in eight numeric characteristics (mercury level, number offish, elevation, surface area, depth, drainage area, runoff factor, and flushing rate). Because not much is known about the distribution of the values we want to compare a lake's value to both the mean and median of all other lakes in the data set with the same value of lake type (LT) and dam status (DAM). The first step is to use PROC MEANS to calculate the mean and median for each of the remaining numeric variables (all eight of them) separately for each combination of LT and DAM. Lakes that don't have a value of LT or DAM reported need to be included in this analysis. I need to use ODS to save the results to a data set named STATZ and ensure the results are not printed to any other destination. Here is my code to do this task and I believe its correct, but just wanted to confirm from you guys. Proc means data = fish ; class lt/missing; class dam/missing; var hg n elv sa z da rf fr ; output out = statz1 mean = median = /autoname; run; The next step is to combine these summary statistics with the original data set using an appropriate data step merge/join technique. This new data set is FISH2. Im confused here as Im not sure on what exactly I have to join summary dataset and original one on? The final step, after the data sets are joined, but in the same DATA step , I need to do the following using arrays. Compute the following for each 8 variables. { Difference from the mean [variable - mean.of.variable] { Percent difference from the mean [(variable - mean.of.variable)/mean.of.variable] { Difference from the median [variable - median.of.variable] { Percent difference from the median [(variable - median.of.variable)/median.of.variable]

devsas · ‎10-13-2017

Ok, I thought I would solve this easily, but after trying various methods im still not getting it. I use the following code to import csv file which has 6 date fields in original date time format stored as character - Some of the values are missing and denoted by 'NULL'. The date values look like this 2017-10-31 00:00:00.000 DATA chpModel; INFILE "C:\Desktop\CHP model results\CHP model results 12 Oct.csv" DELIMITER = "," MISSOVER DSD FIRSTOBS=2 LRECL=32767; Length acctno $50 subsno $50 phaseid $50 AcctnoActiveMonth $50 PhaseActiveMonth $50 PhaseStart $50 Phaseend $50 LastmonthBenpkg $20 TotalPremium Totalpayment Balance AR DeferredRevenue Grace_period_Term $50 Financial_term $50 ; INPUT acctno $ subsno $ phaseid $ AcctnoActiveMonth $ PhaseActiveMonth $ PhaseStart $ Phaseend $ LastmonthBenpkg $ TotalPremium Totalpayment Balance AR DeferredRevenue Grace_period_Term $ Financial_term $ ; format AcctnoActiveMonth mmddyy10 PhaseActiveMonth mmddyy10 PhaseStart mmddyy10 Phaseend mmddyy10 ; RUN; The file is imported fine but the formatting stays the same as original. I also tried importing dates as numeric, but the same result. I know one of the things I can do is converting to date from character once the import is done. But for one, it will change the sequence of the fields and second, with the formatting I have, I was having problem even doing that. I want the final dates to look like 10/31/2017

devsas · ‎09-22-2017

Another way if you prefer proc sql- proc sql; create table want as select id, max(flag1) as flag1, max(flag2) as flag2, max(flag3) as flag3, max(flag4) as flag4 from have group by id; quit;

devsas · ‎09-21-2017

Thanks Reeza. One thing I noticed in your code was having additonal put inside the input function- sas_date=input(put(date1, 8.), yymmdd8.); But if I remove that and use just sas_date=input(date1, yymmdd8.); it works fine too. So, not sure whats the purpose of put function here?

devsas · ‎09-21-2017

No, they are all 7 digits from what I know. They are formatted as number though.

devsas · ‎09-21-2017

Good Afternoon everyone! So, I have got these dates which look like these:- 1980901, 1990831, 2011011, 2130801 but which mean actually 19980901, 19990831, 20011011, 20130801 Basically you need to replace first digit by 19 if the first digit is 1 and by 20 if the first digit is 2. I tried the following code, and not sure what im doing incorrect. data test; input date_ ; cards; 1980901 1990831 2011011 2130801 ; run; data want; set test ; LENGTH DATE1 $12 ; if substr(date_, 1, 1) = 1 then date1 = 19||substr(date_, 2, 6) ; ELSE IF substr(date_, 1, 1) = 2 then date1 = 20||substr(date_, 2, 6) ; RUN;

devsas · ‎09-18-2017

Thanks so much! I used the code in the link you provided replacing relevant names for database and server names. It seems to work, except that i was not sure of what to put for schema and removing that seems to work. But, just want to confirm if schema option is optional? libname mydata odbc complete="driver=SQL Server; database=DHS; server=EIMSQLCLS2X\ODE" ;

devsas · ‎09-18-2017

Good Morning everyone! So, Im trying to connect with SQL server using connect to ODBC and it works well as long as im quering the data from a table in SQL server. But, at times, I need to join the table im quering with a local SAS dataset. I use the following code and it gives me error. Ofcourse there is an option of getting the whole table from sql sever first and then do the join, but the tables im working with are very big tables in sql server and also the data changes in them constantly, so bringing them first to SAS will be a pain. In the code below, enbmast is the big table in sql server. The code below works fine if I dont join it with members table. proc sql; connect to odbc as AX (dsn="EIM"); create table test_ as select * from connection to AX ( select a.acctno, SUM(a.trnamt) AS TRNAMT1, b.memID from enbmast a join work.members b on a.acctno =b.acctno GROUP BY ACCTNO ); quit; Thanks so much in advance!

Online Status	Offline
Date Last Visited	‎10-04-2019 10:53 PM

Re: Flagging records based on a criteria

Re: Flagging records based on a criteria

Re: Flagging records based on a criteria

Re: Flagging records based on a criteria

Re: Flagging records based on a criteria

Flagging records based on a criteria

How to pick records for ID's which dont have certain codes

Re: Extracting all records for only two unique variable values

Extracting all records for only two unique variable values

Re: How to filter for only one record per provider number

Re: How to pick records for ID's which dont have certain codes

Re: How to pick records for ID's which dont have certain codes

Re: How to filter for only one record per provider number

Re: How to filter for only one record per provider number

Re: Displaying data range in proc sql summary

How to pick records for ID's which dont have certain codes

Re: How to filter for only one record per provider number

Re: Summary of flags

Calculating driving distances based on addresses

Re: counting distinct values over a range of variables

Re: Proc means summary statistics, join etc

Re: Proc means summary statistics, join etc

Re: Proc means summary statistics, join etc

Re: Proc means summary statistics, join etc

Re: Proc means summary statistics, join etc

Re: Proc means summary statistics, join etc

Re: Proc means summary statistics, join etc

Proc means summary statistics, join etc

importing CSV file with different datetime format

Re: Summary of flags

Re: converting 7 digit date to regular date

Re: converting 7 digit date to regular date

converting 7 digit date to regular date

Re: Connect to ODBC and using join with existing SAS dataset

Connect to ODBC and using join with existing SAS dataset