It’s a simple question really…why is the default memsize option 2G when realistically if you’re working w/any sizeable dataset or doing memory intensive procedures like multiple imputation you will need way more memory than this? Also, realistically (hopefully) most ppl are operating systems w/at least 16G of RAM, couldnt the default memsize reasonably be increased to at LEAST 4G if not bigger? I just set mine to max, but wondering why default is so low. ... View more

I don't know if I am just struggling to identify a simple solution to this problem, but I have a dataset (in wide format) which has one record for unique participants within the dataset.   The dataset is wide across siblings, in the sense that each unique identifier has one or more siblings identified by their own unique identifier, and these siblings may or may not also themselves be in the dataset as a unique participant.   each sibling is identified in the dataset by their own unique id, and as shown in the picture below, any missing value indicates no sibling (past the last non-missing value, across 14 siblings (the max in the dataset) You'll notice just from this incomplete screenshot of the dataset which is significantly larger, that some of the sibling unique identifiers (eg (sib01_uniqid, sib02_uniqid, etc) are also represented in the unique identifier field for the original observations (uniqid), as well as some that never appear as a simple unique ID in the dataset (e.g., they are known within the broader dataset, but not included in this specific subsample).   Is there an easy/simple way for me to identify those individuals who do have siblings among the original ID's, and create an indicator as such? I am sure there must be a relatively simple way, either in long or in wide format - I have both formats of this dataset available.   Any help is greatly appreciated! ... View more

Hello all,   I have a somewhat complex conditional delete that I am trying to figure out, and it is based on repeated observations, and I have three different types of repeaters that need to be pruned. Essentially, I have a dataset with people who have been interviewed at 2 times.  I want to delete observations for all of these repeaters so that they only have one visit in the dataset, and I would like to do this conditionally on two variables (ACECAT, which is ACEs category, and INT_D, which is interview date). There are three types of repeaters that I need to remove one visit from. They can be identified as repeaters via the use of another variable, site_ID (though in this dataset I've actually restricted so that only repeaters are present). The first type of repeater is missing ACECAT at BOTH visits - for these, it doesn't really matter which observation gets deleted, but for the sake of consistency I would like to delete the first observation and keep the last. INT_D is formatted in DDMMMYYYY format.  the second type of repeater has answered ACECAT at one visit and is missing ACECAT in the other visit - for these types of repeaters, I want to drop the observation where ACECAT is missing.  the third type of repeater has answered ACECAT at both visits. For these types of repeaters, I want to keep only the most recent observation and drop the last observation.  I've been trying to think of a data step (or multiple data steps if needed) that can solve this problem, but am drawing a blank. Can anyone help? ... View more