Hi,   We have recently updated to EG 7.15 HF1 (7.100.5.6107) from EG 7.15 and are hitting a new error.   Whenever we attempt to save a project to our shared network location we get the following error message. "An error occurred saving the file. Access to the path is denied."   Saving projects locally to the C drive seem to be unaffected but saving to our main shared network location hits the error. When saving over an existing project we get the error and nothing happens, using 'save as' creates a new project object in the location but it is 0kb and EG says it is not a valid file on opening.   Is this an issue anyone else has come across?   Tim ... View more

Hi all,   I am hitting a slight bug with prompts in Enterprise guide 7.11 HF3 (7.100.1.2805).   I have built a reporting process for our users in EG where they select a week to report on in a prompt. The prompt is built as shown in the image below, as a date prompt where users enter a value with the date type set to week.     When the process using the prompt is run, enterprise guide sets some macro variables, prompt_1 and prompt_1_end as date9 format dates for the first and last date from the week selected.   The bug I am getting is that, generally the dates selected will be Mon-Sun for the week selected except for when week 1 of any year is selected where, the prompt examples show the correct dates but the actual values are Thu-Wed. This behaviour seems to be true for most years not just 2016.   e.g. Week53 2015 = 28DEC2015 - 03JAN2016 Week01 2016 = 07JAN2016 - 13JAN2016 (should be 04JAN2016-10JAN2016) Week02 2016 = 11JAN2016 - 17JAN2016   Screen shots of prompt selecting and log after running:     Has anyone else come across this or found a solution? Obviously it won't be difficult to change my process to select a single date and then calculate the week required myself but if i can just fix EG it will probably be less work...   Tim ... View more

Hello,   I need some help to find the fastest way to get the next ID number from a large dataset.   I am building a process to run several hundred forecast scenarios at a time in SAS and have a dataset acting as a dimension table which contains an ID number which is used to join the Dimension and Fact tables together in later processes. When each individual process is run to create a new scenario I need to look up the next ID number from my DIM table to use when creating the new scenario then I use proc append to add the new DIM and FACT records to the appropriate tables.   To look up the next ID number I was using the following process. proc sql noprint; select sum(max(SCENARIO_ID),1) into :l_next_forecast_id from LIB.DIM_TABLE; quit;   When stress testing with more data(DIM table has over 5 million records) this query slowed down to over 2 seconds per scenario as it is having to read the entire table each time, so I started using the following process instead. data _null_; set LIB.DIM_TABLE nobs=NOBS point=NOBS; call symputx('l_next_forecast_id',sum(SCENARIO_ID,1)); stop; run; This is fine if I can assume that nothing will ever go wrong to make the last record not contain the max value or no data is ever deleted however paranoia tells me otherwise, so I would need to sort the data beforehand which again added too much time to the process. This was surprising because, given that the data in the DIM table was already sorted and I am appending a new record(also sorted with proc sort) with a higher ID number I would expect from the documenation below that the sort indicator would be retained by the proc append, meaning that the sort step should only fire if something had gone wrong and flag that no sort required if everything was fine. Proc Append Docs The documented behaviour works as expected when dealing with work datasets but doesn't work when I am appending to a SAS/SHARE library so without the sort indicator retained the proc sort always fires and slows down the process. I could use proc datasets to add the sort indicator back onto the table after the proc append but this really defies the point since if something has gone wrong and the append had happended incorrectly I don't want to override it so it stays broken for next time.   The last thing I tried was adding an index to the table and using proc contents with the centiles option to read the max key value however keeping the index on the dataset made the append steps too slow so i'm back to square 1.   For reference, an example of the process I am currently using where the sort indicator is lost is as follows. As i said before, this process works when the PLAYPEN library is assigned normally rather than using SAS/SHARE but the final library will need to use SAS/SHARE so changing that isn't an option. Libname PLAYPEN '/data/dtf/playpen' server=sshare3 sapw=XXXXXXXX; data PLAYPEN.SET1; set sashelp.class; ORDER_VARIABLE=1; run; proc sort data=PLAYPEN.SET1; by ORDER_VARIABLE; run; data WORK.SET2; set sashelp.class; ORDER_VARIABLE=2; run; proc sort data=WORK.SET2; by ORDER_VARIABLE; run; /* Before append Sort indicator present */ proc contents data=PLAYPEN.SET1; run; proc append base=PLAYPEN.SET1 data=WORK.SET2; run; /* After append Sort indicator lost */ proc contents data=PLAYPEN.SET1; run;   Does anyone know of a more efficient way to get the next ID number or why my proc append steps would be dropping the sort indicator when using SAS/SHARE libraries?   Tim ... View more