我已经查看过一篇关于它的帖子(SPSS Last available measurement in a row of variables)，但我仍然对如何更有效地执行此任务心存疑虑。我有一个包含 9800 万行和 120 个变量的数据集(每个月一个，从 2005 年 1 月到 2014 年 12 月)。 我需要为数据集中的每个观察值确定一行变量中的第一个和最后一个有效(非缺失)测量值。 数据集如下所示:
v1 v2 ... v120
1 2 ... 5
. 2 ... 5
3 1 ... .
我已经尝试过使用循环的版本(也建议在这个主题上:SPSS Last available measurement in a row of variables)。我使用了下面的语法，但它们不起作用。我不断收到错误消息……很可能是因为我不理解其中的所有步骤，所以误用了它。
DEFINE LAST_VALID ()
!DO !@ = 1 !TO 120 .
!LET !a = !CONCAT("v", !@) .
COMPUTE LAST_VALID = !a .
!DOEND .
!ENDDEFINE.
LAST_VALID .
EXECUTE.
错误消息(其中一些):
Error # 4382 in column 1024. Text: (End of Command) >An equals sign was >not found when expected after a target variable in a >COMPUTE command. >Execution of this command stops.
Warning # 231 >The depth of macro nesting has reached the current limit. >To increase the >limit, use SET MNEST. To check the limit use the SHOW >command.
Warning # 210 in column 9. Text: !ERROR_MACRO >A macro symbol is invalid >in this context. >The symbol will be treated as an invalid special >character.
Error # 4285 in column 9. Text: !ERROR_MACRO >Incorrect variable name: >either the name is more than 64 characters, or it is >not defined by a >previous command. >Execution of this command stops.
DEFINE FIRST_VALID ()
!DO !@ = 1 !TO 120 .
!LET !a = !CONCAT("v", !@) .
LOOP IF MISSING (FIRST_VALID) = 1.
COMPUTE FIRST_VALID = !a .
END LOOP IF FIRST_VALID > 0.
!DOEND .
!ENDDEFINE.
FIRST_VALID.
EXECUTE.
错误消息(其中一些):
Warning # 231 >The depth of macro nesting has reached the current limit. >To increase the >limit, use SET MNEST. To check the limit use the SHOW >command.
Warning # 210 in column 18. Text: !ERROR_MACRO >A macro symbol is invalid >in this context. >The symbol will be treated as an invalid special >character.
Error # 4007 in column 18. Text: !ERROR_MACRO >The expression is >incomplete. Check for missing operands, invalid operators, >unmatched >parentheses or excessive string length. >Execution of this command stops.
Error # 4846 in column 18. Text: !ERROR_MACRO >The LOOP command contains >unrecognized text after the end of the IF clause.
Warning # 210 in column 9. Text: !ERROR_MACRO >A macro symbol is invalid >in this context. >The symbol will be treated as an invalid special >character.
Error # 4285 in column 9. Text: !ERROR_MACRO >Incorrect variable name: >either the name is more than 64 characters, or it is >not defined by a >previous command. >Execution of this command stops.
Error # 4014 in column 13. Text: !ERROR_MACRO >SPSS Statistics was >expecting an expression but encountered the end of the >command. Check the >expression for omitted or extra operands, operators, and >parentheses. >Execution of this command stops.
Error # 4045. Command name: END LOOP >The END LOOP command does not follow >an unclosed LOOP command. Maybe the LOOP >command was not recognized >because of an error. Use the level-of-control >shown to the left of the >SPSS Statistics commands to determine the range of >LOOPs and DO IFs.*
我可能忘记为某些东西调整我的语法，但我不知道是什么。 如果太明显，我深表歉意……
最佳答案
您的描述和代码片段都不清楚您想要什么。最后一个实际值或该值的索引位置。例如。如果最后一个有效值在 V110 中并且等于 5 记录 100 万，你想知道“110”还是“5”。
这是一个使用 DO REPEAT 的简单方法，它会返回任何一个。 LastVal 在我的例子中是“5”，LastId 是 110。
DO REPEAT V = V1 TO V120 /#i = 1 TO 120.
DO IF NOT MISSING(V).
COMPUTE LastVal = V.
COMPUTE LastId = #i.
END IF.
END REPEAT.
要同时返回第一个索引和值，您可以在 DO REPEAT 中执行第二个 DO IF。
NUMERIC FirstVal FirstId.
DO REPEAT V = V1 TO V120 /#i = 1 TO 120.
DO IF NOT MISSING(V).
COMPUTE LastVal = V.
COMPUTE LastId = #i.
DO IF MISSING(FirstVal).
COMPUTE FirstVal = V.
COMPUTE FirstId = #i.
END IF.
END IF.
END REPEAT.
有 9800 万条记录，这可能需要一段时间 - 特别是如果不在服务器上。您可以使用 VARSTOCASES 进行试验，默认情况下它会删除丢失的数据。但在完成之前，这些会一直持续下去。
您可以通过使用 LOOP 来达到相同的效果，但在遇到第一个有效值时通过中断来提高效率。因此，对于您可以执行的第一个有效值。
VECTOR V = V1 TO V120.
LOOP #i = 1 TO 120.
END LOOP IF NOT MISSING(V(#i)).
COMPUTE FirstVal = V(#i).
COMPUTE FirstId = #i.
对于最后的值，您只需要反转循环即可。
VECTOR V = V1 TO V120.
LOOP #i = 120 TO 1 BY -1.
END LOOP IF NOT MISSING(V(#i)).
COMPUTE LastVal = V(#i).
COMPUTE LastId = #i.
关于SPSS - 如何识别一行 120 个变量中的第一个和最后一个可用测量值，我们在Stack Overflow上找到一个类似的问题： https://stackoverflow.com/questions/36017392/