约束具有相同ID的每个观察值

问题描述

我有一个数据集，在其中工作了一段时间以将其从宽格式转换为长格式。

我们正在跟踪约1.000例1-5个动脉瘤（可能有1个以上的动脉瘤）的患者，并且部分或全部患者接受了不同的治疗方法。该患者可能患有两种动脉瘤，其中一种接受治疗A，另一种接受治疗B。

以下是数据示例：

* Example generated by -dataex-. To install: ssc install dataex
clear
input str32 record_id float treatmentChoice_ byte(treatment_1 treatment_3) float aneurysm_id
"007128de18ce5cb1635b8f27c5435ff3" . . . 1
"007128de18ce5cb1635b8f27c5435ff3" . . . 2
"00abd7bdb6283dd0ac6b97271608a122" . 2 . 1
"00abd7bdb6283dd0ac6b97271608a122" . . . 2
"0142103f84693c6eda416dfc55f65de1" . 1 . 1
"0142103f84693c6eda416dfc55f65de1" . . . 2
"0153826d93a58d7e1837bb98a3c21ba8" . . . 1
"0153826d93a58d7e1837bb98a3c21ba8" . . . 2
"01c729ac4601e36f245fd817d8977917" . 1 . 1
"01c729ac4601e36f245fd817d8977917" . . . 2
"01dd90093fbf201a1f357e22eaff6b6a" . . . 1
"01dd90093fbf201a1f357e22eaff6b6a" . 1 . 2
"0208e14dcabc43dd2b57e2e8b117de4d" . . . 1
"0208e14dcabc43dd2b57e2e8b117de4d" . 1 . 2
"0210f575075e5def7ffa77530ce17ef0" . . . 1
"0210f575075e5def7ffa77530ce17ef0" . . . 2
"022cc7a9397e81cf58cd9111f9d1db0d" . . . 1
"022cc7a9397e81cf58cd9111f9d1db0d" . . . 2
"02afd543116a22fc7430620727b20bb5" . 2 . 1
"02afd543116a22fc7430620727b20bb5" . . . 2
"0303ef0bd5d256cca1c836e2b70415ac" . . . 1
"0303ef0bd5d256cca1c836e2b70415ac" . 1 . 2
"041b2b0cac589d6e3b65bb924803cf1a" . . . 1
"041b2b0cac589d6e3b65bb924803cf1a" . . . 2
"0536317a2bbb936e85c3eb8294b076da" . . . 1
"0536317a2bbb936e85c3eb8294b076da" . 1 . 2
"06161d4668f217937cac0ac033d8d199" . . . 1
"06161d4668f217937cac0ac033d8d199" . . . 2
"065e151f8bcebb27fabf8b052fd70566" . . 1 1
"065e151f8bcebb27fabf8b052fd70566" . . . 2
"065e151f8bcebb27fabf8b052fd70566" . . . 3
"065e151f8bcebb27fabf8b052fd70566" . . . 4
"07196414cd6bf89d94a33e149983d102" . . . 1
"07196414cd6bf89d94a33e149983d102" . . . 2
"0721c38f8275dab504fc53aebcc005ce" . . . 1
"0721c38f8275dab504fc53aebcc005ce" . . . 2
"0721c38f8275dab504fc53aebcc005ce" . . . 3
"0721c38f8275dab504fc53aebcc005ce" 1 . . 4
"07bef516d53279a3f5e477d56d552a2b" . . . 1
"07bef516d53279a3f5e477d56d552a2b" . 2 . 2
"08678829b7e0ee6a01b17974b4d19cfa" . . . 1
"08678829b7e0ee6a01b17974b4d19cfa" . . . 2
"08bb6c65e63c499ea19ac24d5113dd94" . . . 1
"08bb6c65e63c499ea19ac24d5113dd94" . . . 2
"08f036417500c332efd555c76c4654a0" . . . 1
"08f036417500c332efd555c76c4654a0" . . . 2
"090c54d021b4b21c7243cec01efbeb91" . . . 1
"090c54d021b4b21c7243cec01efbeb91" . . . 2
"09166bb44e4c5cdb8f40d402f706816e" . . . 1
"09166bb44e4c5cdb8f40d402f706816e" . 1 . 2
"0930159addcdc35e7dc18812522d4377" . . . 1
"0930159addcdc35e7dc18812522d4377" . . . 2
"096844af91d2e266767775b0bee9105e" . . . 1
"096844af91d2e266767775b0bee9105e" . 2 . 2
"09884af1bb9d59803de0c74d6df57c23" . . . 1
"09884af1bb9d59803de0c74d6df57c23" . 2 . 2
"09e03748da35e9d799dc5d8ddf1909b5" . . . 1
"09e03748da35e9d799dc5d8ddf1909b5" . . . 2
"0a4ce4a7941ff6d1f5c217bf5a9a3bf9" . . . 1
"0a4ce4a7941ff6d1f5c217bf5a9a3bf9" . . . 2
"0a5db40dc58e97927b407c9210aab7ba" 4 . . 1
"0a5db40dc58e97927b407c9210aab7ba" . . . 2
"0a73c992955231650965ed87e3bd52f6" . . . 1
"0a73c992955231650965ed87e3bd52f6" . 2 . 2
"0a84ab77fff74c247a525dfde8ce988c" 1 . 2 1
"0a84ab77fff74c247a525dfde8ce988c" . . . 2
"0a84ab77fff74c247a525dfde8ce988c" . . . 3
"0af333ae400f75930125bb0585f0dcf5" . . . 1
"0af333ae400f75930125bb0585f0dcf5" . . . 2
"0af73334d9d2166191f3385de48f15d2" . 1 . 1
"0af73334d9d2166191f3385de48f15d2" . . . 2
"0b341ac8f396a8cdb88b7c658f66f653" . . . 1
"0b341ac8f396a8cdb88b7c658f66f653" . . . 2
"0b35cf4beb830b361d7c164371f25149" . 1 . 1
"0b35cf4beb830b361d7c164371f25149" . . . 2
"0b3e110c9765e14a5c41fadcc3cfc300" . . . 1
"0b6681f0f441e69c26106ab344ac0733" . . . 1
"0b6681f0f441e69c26106ab344ac0733" . . . 2
"0b8d8253a8415275dbc2619e039985bb" 4 . . 1
"0b8d8253a8415275dbc2619e039985bb" . . . 2
"0b8d8253a8415275dbc2619e039985bb" . . . 3
"0b92c26375117bf42945c04d8d6573d4" . 2 . 1
"0b92c26375117bf42945c04d8d6573d4" . . . 2
"0ba961f437f43105c357403c920bdef1" . . . 1
"0ba961f437f43105c357403c920bdef1" . . . 2
"0bb601fabe1fdfa794a5272408997a2f" . . . 1
"0bb601fabe1fdfa794a5272408997a2f" . . . 2
"0c75b36e91363d596dc46bd563c3f5ef" . 1 . 1
"0c75b36e91363d596dc46bd563c3f5ef" . . . 2
"0d461328a3bae7164ce7d3a10f366812" . . . 1
"0d461328a3bae7164ce7d3a10f366812" . 2 . 2
"0d4cc4eb459301a804cbef22914f44a3" . 1 . 1
"0d4cc4eb459301a804cbef22914f44a3" . . . 2
"0d4e29e11bb94e922112089f3fec61ef" . . . 1
"0d4e29e11bb94e922112089f3fec61ef" . 1 . 2
"0d513c74d667f55c8f4a9836c304149c" . 1 . 1
"0d513c74d667f55c8f4a9836c304149c" . . . 2
"0da25de126bb3b3ee565eff8888004c2" . . . 1
"0da25de126bb3b3ee565eff8888004c2" . 1 . 2
"0db9ae1f2201577f431b7603d0819fa6" . . . 1
end
label values treatment_1 treatment_1_
label def treatment_1_ 1 "Observation",modify
label def treatment_1_ 2 "Afsluttet",modify
label values treatment_3 treatment_3_
label def treatment_3_ 1 "Observation",modify
label def treatment_3_ 2 "Afsluttet",modify

如您所见，在此示例中，存在三种不同的处理方式，并且我已按record_ID（患者）对观察结果进行了排序。请注意，每个患者（record_ID）可以出现多次。实际上，我已经扩展了数据集，因此，如果患者有4个动脉瘤，那么将有4个观察结果，因为统计数据基于动脉瘤而不是患者。

我的问题是，这些观察值中的哪一个似乎描述每个动脉瘤得到的治疗似乎是随机的，我想添加一个变量treatment，其中列出了相应动脉瘤ID的治疗。还要注意，treatmentChoice_的意思是“动脉瘤1得到了哪种治疗？” treatmentChoice_1的意思是“动脉瘤2接受了哪种治疗？”

有没有办法说：

“对于每个相同的record_ID，如果动脉瘤ID为1，则浏览treatmentChoice_并将treatment设置为该值。然后对treatmentChoice_1，treatmentChoice_3执行相同的操作如果动脉瘤ID分别为2或3，则将treatment设置为其值。

解决方法

如果我正确地遵循了此操作，则您希望从每个观察值的一些变量中选择一个非缺失值。为此，您可以使用max()中的min()或rowmin()或rowmax()或egen函数。

有了您的示例数据（谢谢），我明白了。请注意两个未标记的值4。

. generate treatment = max(treatmentChoice_,treatment_1,treatment_3)
(73 missing values generated)

. label val treatment treatment_1_

. tab treatment

  treatment |      Freq.     Percent        Cum.
------------+-----------------------------------
Observation |         16       59.26       59.26
  Afsluttet |          9       33.33       92.59
          4 |          2        7.41      100.00
------------+-----------------------------------
      Total |         27      100.00