Due to the ( lack of ) precision in floating point operations on the computer , the value val3 is nor excactly -600 in both cases. insert valx=val3+600; put valx= 20.15; in each data step and you will see valx=0.000000000000000 and valx=0.000000000000114 and since the format is separated on either side of -600 , the results differ. At least on mi Intel based computer. ... View more

You can use proc datasets to see what tables are in a library. I often find it usefull to change the "name" for any table to "library: name " in order to see what library any given table actually belongs to. The name is just some metadata and is not used to identify any objects and any change in name will automatically be reflected in all visual presentations of the table. ( The name of the physical table will not change.) Henrik ... View more