===== Whats new 08nov2012 ===== 1. Date corrections: in Stata 10 only // Stefan Gawrich 2. Characters other than 'A'..'Z','a'..'z','0'..'9','_' in variable names are replaced with '_' to produce valid Stata names. 3. Extended list of exceptional variable names: byte, long, etc. 4. Show to user, which renames took place: SPSS_Name ---> Stata_Name 5. Variables formatting: g-->f, 12-->16, NOTE rounding problems for 0.5! 6. Values alignment: left/right. 7. Added links in output for #1 and #4. ===== Whats new 19sep2010 ===== 0. 64-bit support! (works in 64-bit Stata, does not work with 2GB+ datasets). 1. Original variable names from SPSS are saved into variables' characteristics 2. Value labels for numeric missing values [ // suggested by Dirk Enzmann] Original numeric values are prepended to the labels, e.g.: .a [0]IAP .b [8]Dont Know .c [9]No Answer Original labels are left available for reference, even though the original missing values (0,8,9) were recoded to Stata's extended missings: ---> 0 IAP 1 COULD HAVE BEEN INTERVIEWED IN ENGLISH 2 WOULD HAVE BEEN EXCLUDED AS LANGUAGE PROBLEM ---> 8 Dont Know ---> 9 No Answer .a [0]IAP .b [8]Dont Know .c [9]No Answer 3. Long strings (248+ were a problem, not anymore) // Christina Tippman 4. Terminators were not correctly written for labels spanning across the Stata's limits // Gunilla Pettersson, Hua Peng 5. Labelling of string variables caused problems // currently the problem may show up in the case when the same 8-bytes value // is labelled in a numeric (first) and a string (subsequent) variables 6. If the file is locked (in use) - do not crash anymore. 7. If memory allocation is denied - do not crash anymore. 8. Last variable was not renamed appropriately. ~ Aki Kia? 9. Missing values for strings confused the program. They are now tolerated. 10. Some files declared holding -1 cases. Alternative case count is applied to such files. 11. Some labelled integers can't hold their labels in Stata, e.g.: 9999999998, program crashed before, not anymore. 12. Support of embedded document - document lines are printed on the screen. -------- PLANNED, BUT NOT IMPLEMENTED: ---------------------------------------- 1. Dates corrections 2. Variable "FILTER_$" to be used as a default filter 3. Variable "WEIGHT" to be used to set up svy settings 4. Not investigated behaviour when labels are applied to values falling into the range of missing values (in Stata they will all have the same value, e.g. .c, but what will happen to the labels?).