附加数据会产生错误的结果

Question

我有两个不同年份的数据集,我试图使用该append命令进行组合 .两者都包括许多变量和一个识别变量的人,但这些变量并不相同.

数据集1:

PID        Year     Car  Sex       Age
420201     2016     0    Female    70
420202     2016     0    Male      87
420204     2016     0    Female    62
420205     2016     1    Female    34
420207     2016     1    Male      48

数据集2:

PID        Year     Car   Sex      Age
420202     2014     1     Male     59
420204     2014     0     Female   76
420205     2014     1     Male     37
420207     2014     1     Male     23

问题在于,当我尝试附加这些数据集时,Stata会生成一个数据集,其中一个数据集中某些标识符的值被错误地赋予其他数据集的标识符.

附加数据集:

PID        Year     Car   Sex      Age
420201     2016     0     Female   70
420201     2014     1     Male     59
420202     2016     0     Male     87
420202     2014     0     Female   76
420204     2016     0     Female   62
420204     2014     1     Male     37
420205     2016     1     Female   34
420205     2014     1     Male     23
420207     2016     1     Male     48
420207     2014     1     Male     23

有没有解决这个问题？

Answer 1

我之前遇到过这个问题并且它发生了,因为PID你看到的'值'实际上是附加到值的标签1, 2, 3, 4, 5.

为了说明这一点,请考虑以下示例:

clear

input PID Year Car str6 Sex Age
1 2016 0 Female 70
2 2016 0 Male 87
3 2016 0 Female 62
4 2016 1 Female 34
5 2016 1 Male 48
end

label define PID 1 "420201" 2 "420202" 3 "420204" 4 "420205" 5 "420207"
label values PID PID

list
     +------------------------------------+
     |    PID   Year   Car      Sex   Age |
     |------------------------------------|
  1. | 420201   2016     0   Female    70 |
  2. | 420202   2016     0     Male    87 |
  3. | 420204   2016     0   Female    62 |
  4. | 420205   2016     1   Female    34 |
  5. | 420207   2016     1     Male    48 |
     +------------------------------------+

list, nolabel
     +---------------------------------+
     | PID   Year   Car      Sex   Age |
     |---------------------------------|
  1. |   1   2016     0   Female    70 |
  2. |   2   2016     0     Male    87 |
  3. |   3   2016     0   Female    62 |
  4. |   4   2016     1   Female    34 |
  5. |   5   2016     1     Male    48 |
     +---------------------------------+

因此,当您尝试append以下情况时:

clear

input PID Year Car str6 Sex Age
1 2014 1 Male 59
2 2014 0 Female 76
3 2014 1 Male 37
4 2014 1 Male 23
end

label define PID 1 "420202" 2 "420204" 3 "420205" 4 "420207"
label values PID PID

save data2, replace

append using data1

sort PID

list

     +------------------------------------+
     |    PID   Year   Car      Sex   Age |
     |------------------------------------|
  1. | 420202   2014     1     Male    59 |
  2. | 420202   2016     0   Female    70 |
  3. | 420204   2016     0     Male    87 |
  4. | 420204   2014     0   Female    76 |
  5. | 420205   2014     1     Male    37 |
     |------------------------------------|
  6. | 420205   2016     0   Female    62 |
  7. | 420207   2016     1   Female    34 |
  8. | 420207   2014     1     Male    23 |
  9. |      5   2016     1     Male    48 |
     +------------------------------------+

您的价值标签可能有不同的定义,但想法是一样的.

为了能够append正确地处理这两个数据集,首先需要转换PID为字符串:

foreach dta in data1 data2 {
    use `dta', clear
    decode PID, generate(PID2)

    drop PID
    rename PID2 PID
    save `dta', replace
}

append using data1

order PID
sort PID

list

     +------------------------------------+
     |    PID   Year   Car      Sex   Age |
     |------------------------------------|
  1. | 420201   2016     0   Female    70 |
  2. | 420202   2016     0     Male    87 |
  3. | 420202   2014     1     Male    59 |
  4. | 420204   2016     0   Female    62 |
  5. | 420204   2014     0   Female    76 |
     |------------------------------------|
  6. | 420205   2016     1   Female    34 |
  7. | 420205   2014     1     Male    37 |
  8. | 420207   2016     1     Male    48 |
  9. | 420207   2014     1     Male    23 |
     +------------------------------------+

您可能还希望使用该destring命令将新字符串PID变量转换为数字变量.