根据另一列的条件向行添加索引 - R答案

【问题标题】：Adding an index to a row based on a condition of another column - R根据另一列的条件向行添加索引 - R
【发布时间】：2019-12-19 01:23:58
【问题描述】：

我有来自自动电话调查的数据。此调查包括 2 次呼叫者断开连接时的回拨。

这是来自一位受访者的数据。不过，我有数千个类似的数据需要处理。

第一列是受访者的ID，第二列是订购ID（因为时间戳不是毫秒级的），第三列是调查工具和受访者手机（本质上是受访者）之间的交互，第四列是日期时间戳，最后是我之前计算的行之间的时间差。



test <- data.frame(matrix(c(
111,    2340,   'Enqueueing call',  '12/2/19 14:53:57', NA,
111,    6174,   'Call expired, will be retried in next schedule window',    '12/2/19 20:14:22', 19225,
111,    10386,  'Answer',   '12/3/19 15:48:56', 70474,
111,    10387,  'Contacted',    '12/3/19 15:48:56', 0,
111,    10388,  'Intro',    '12/3/19 15:48:56', 0,
111,    10389,  'Timeout',  '12/3/19 15:49:16', 20,
111,    10390,  'Intro',    '12/3/19 15:49:16', 0,
111,    10391,  '1',    '12/3/19 15:49:30', 14,
111,    10392,  'Started',  '12/3/19 15:49:30', 0,
111,    10393,  'Q1',   '12/3/19 15:49:30', 0,
111,    10394,  '1',    '12/3/19 15:49:45', 15,
111,    10395,  'Q2',   '12/3/19 15:49:45', 0,
111,    10396,  'Timeout',  '12/3/19 15:49:54', 9,
111,    10397,  'Q2',   '12/3/19 15:49:54', 0,
111,    10398,  'Timeout',  '12/3/19 15:50:03', 9,
111,    10399,  'Q2',   '12/3/19 15:50:03', 0,
111,    10400,  'Timeout',  '12/3/19 15:50:11', 8,
111,    17658,  'Timeout. Call failed.',    '12/4/19 17:50:27', 93616,
111,    19932,  'Call expired, will be retried in next schedule window',    '12/4/19 20:45:17', 10490,
111,    25647,  'Call expired, will be retried in next schedule window',    '12/5/19 20:33:27', 85690,
111,    31516,  'Call expired, will be retried in next schedule window',    '12/6/19 20:17:18', 85431,
111,    36781,  'Call expired, will be retried in next schedule window',    '12/7/19 20:02:16', 85498,
111,    38718,  'Answer',   '12/8/19 10:24:07', 51711,
111,    38719,  'Q2',   '12/8/19 10:24:07', 0,
111,    38720,  'Timeout',  '12/8/19 10:24:16', 9,
111,    38721,  'Q2',   '12/8/19 10:24:16', 0,
111,    38722,  'Timeout',  '12/8/19 10:24:24', 8,
111,    38723,  'Q2',   '12/8/19 10:24:24', 0,
111,    38724,  'Timeout',  '12/8/19 10:24:33', 9,
111,    45339,  'Timeout. Call failed.',    '12/9/19 12:25:08', 93635,
111,    49026,  'Call expired, will be retried in next schedule window',    '12/9/19 20:31:26', 29178,
111,    53972,  'Call expired, will be retried in next schedule window',    '12/10/19 20:14:38',    85392,
111,    58277,  'Answer',   '12/11/19 17:55:29',    78051,
111,    58278,  'Q2',   '12/11/19 17:55:29',    0,
111,    58279,  'Timeout',  '12/11/19 17:55:38',    9,
111,    58280,  'Q2',   '12/11/19 17:55:38',    0,
111,    58281,  'Timeout',  '12/11/19 17:55:46',    8,
111,    58282,  'Q2',   '12/11/19 17:55:46',    0,
111,    58283,  '61',   '12/11/19 17:55:57',    11,
111,    58284,  'Q3',   '12/11/19 17:55:57',    0,
111,    58285,  '4',    '12/11/19 17:56:13',    16,
111,    58286,  'Q4',   '12/11/19 17:56:13',    0,
111,    58288,  '3',    '12/11/19 17:56:42',    29,
111,    58289,  'Interim partial',  '12/11/19 17:56:42',    0,
111,    58290,  'Q5',   '12/11/19 17:56:42',    0,
111,    58291,  '3',    '12/11/19 17:56:59',    17,
111,    58292,  'Q6',   '12/11/19 17:56:59',    0,
111,    58293,  '3',    '12/11/19 17:57:25',    26,
111,    58294,  'Q7',   '12/11/19 17:57:25',    0,
111,    58295,  '3',    '12/11/19 17:57:38',    13,
111,    58296,  'Q8',   '12/11/19 17:57:38',    0,
111,    58297,  '3',    '12/11/19 17:57:50',    12,
111,    58298,  'Q9',   '12/11/19 17:57:50',    0,
111,    58299,  'Timeout',  '12/11/19 17:58:09',    19,
111,    58300,  'Q9',   '12/11/19 17:58:09',    0,
111,    58301,  '10',   '12/11/19 17:58:32',    23,
111,    58302,  'Q10',  '12/11/19 17:58:32',    0,
111,    58303,  '1',    '12/11/19 17:58:49',    17,
111,    58304,  'Q11',  '12/11/19 17:58:49',    0,
111,    58307,  '3',    '12/11/19 17:59:02',    13,
111,    58308,  'Q12',  '12/11/19 17:59:02',    0,
111,    58309,  'Timeout',  '12/11/19 17:59:23',    21,
111,    58310,  'Q13',  '12/11/19 17:59:23',    0,
111,    58311,  'Timeout',  '12/11/19 17:59:44',    21,
111,    58312,  'Q13',  '12/11/19 17:59:44',    0,
111,    58313,  '4',    '12/11/19 17:59:51',    7,
111,    58314,  'Q14',  '12/11/19 17:59:51',    0,
111,    58318,  '2',    '12/11/19 18:00:19',    28,
111,    58319,  'Q15',  '12/11/19 18:00:19',    0,
111,    58320,  '3',    '12/11/19 18:00:36',    17,
111,    58321,  'Q16',  '12/11/19 18:00:36',    0,
111,    58322,  '4',    '12/11/19 18:01:04',    28,
111,    58323,  'Q17',  '12/11/19 18:01:04',    0,
111,    58324,  '1',    '12/11/19 18:01:18',    14,
111,    58325,  'Q18',  '12/11/19 18:01:18',    0,
111,    58327,  '4',    '12/11/19 18:01:42',    24,
111,    58328,  'Q19',  '12/11/19 18:01:42',    0,
111,    58329,  '1',    '12/11/19 18:01:56',    14,
111,    58330,  'Q20',  '12/11/19 18:01:56',    0,
111,    58331,  '1',    '12/11/19 18:02:19',    23,
111,    58332,  'Q21',  '12/11/19 18:02:19',    0,
111,    58333,  '1',    '12/11/19 18:02:28',    9,
111,    58334,  'Q22',  '12/11/19 18:02:28',    0,
111,    58335,  '2',    '12/11/19 18:02:52',    24,
111,    58336,  'Completed',    '12/11/19 18:02:52',    0,
111,    58337,  'Complete Message', '12/11/19 18:02:52',    0,
111,    58338,  'Thank you',    '12/11/19 18:02:52',    0),
nrow=87, ncol=5, byrow=T,
dimnames=list(c(NULL), c("Respondent.ID", "order.ID", "Interaction", "Datetime", "difftime"))))

我需要总结 difftime 的时间，但仅基于某些条件。本质上，我希望数据如下所示，因此我可以总结 Include = 1 的时间。

条件 1：Interaction 之后的行 = "Answer" 和 Interaction 之前的行之间的所有行 = "Timeout. Call failed."应该包括 = 1。条件 2：Interaction = "Answer" 之后的行和 Interaction = "Completed" 之前的行之间的所有行都应该是 Include = 1。



new <- data.frame(matrix(c(
111,    2340,   'Enqueueing call',  '   12/2/19 14:53:57    ',  NA, 0,
111,    6174,   'Call expired, will be retried in next schedule window',    '   12/2/19 20:14:22    ',  19225,  0,
111,    10386,  'Answer',   '   12/3/19 15:48:56    ',  70474,  0,
111,    10387,  'Contacted',    '   12/3/19 15:48:56    ',  0,  1,
111,    10388,  'Intro',    '   12/3/19 15:48:56    ',  0,  1,
111,    10389,  'Timeout',  '   12/3/19 15:49:16    ',  20, 1,
111,    10390,  'Intro',    '   12/3/19 15:49:16    ',  0,  1,
111,    10391,  '1',    '   12/3/19 15:49:30    ',  14, 1,
111,    10392,  'Started',  '   12/3/19 15:49:30    ',  0,  1,
111,    10393,  'Q1',   '   12/3/19 15:49:30    ',  0,  1,
111,    10394,  '1',    '   12/3/19 15:49:45    ',  15, 1,
111,    10395,  'Q2',   '   12/3/19 15:49:45    ',  0,  1,
111,    10396,  'Timeout',  '   12/3/19 15:49:54    ',  9,  1,
111,    10397,  'Q2',   '   12/3/19 15:49:54    ',  0,  1,
111,    10398,  'Timeout',  '   12/3/19 15:50:03    ',  9,  1,
111,    10399,  'Q2',   '   12/3/19 15:50:03    ',  0,  1,
111,    10400,  'Timeout',  '   12/3/19 15:50:11    ',  8,  1,
111,    17658,  'Timeout. Call failed.',    '   12/4/19 17:50:27    ',  93616,  0,
111,    19932,  'Call expired, will be retried in next schedule window',    '   12/4/19 20:45:17    ',  10490,  0,
111,    25647,  'Call expired, will be retried in next schedule window',    '   12/5/19 20:33:27    ',  85690,  0,
111,    31516,  'Call expired, will be retried in next schedule window',    '   12/6/19 20:17:18    ',  85431,  0,
111,    36781,  'Call expired, will be retried in next schedule window',    '   12/7/19 20:02:16    ',  85498,  0,
111,    38718,  'Answer',   '   12/8/19 10:24:07    ',  51711,  0,
111,    38719,  'Q2',   '   12/8/19 10:24:07    ',  0,  1,
111,    38720,  'Timeout',  '   12/8/19 10:24:16    ',  9,  1,
111,    38721,  'Q2',   '   12/8/19 10:24:16    ',  0,  1,
111,    38722,  'Timeout',  '   12/8/19 10:24:24    ',  8,  1,
111,    38723,  'Q2',   '   12/8/19 10:24:24    ',  0,  1,
111,    38724,  'Timeout',  '   12/8/19 10:24:33    ',  9,  1,
111,    45339,  'Timeout. Call failed.',    '   12/9/19 12:25:08    ',  93635,  0,
111,    49026,  'Call expired, will be retried in next schedule window',    '   12/9/19 20:31:26    ',  29178,  0,
111,    53972,  'Call expired, will be retried in next schedule window',    '   12/10/19 20:14:38   ',  85392,  0,
111,    58277,  'Answer',   '   12/11/19 17:55:29   ',  78051,  0,
111,    58278,  'Q2',   '   12/11/19 17:55:29   ',  0,  1,
111,    58279,  'Timeout',  '   12/11/19 17:55:38   ',  9,  1,
111,    58280,  'Q2',   '   12/11/19 17:55:38   ',  0,  1,
111,    58281,  'Timeout',  '   12/11/19 17:55:46   ',  8,  1,
111,    58282,  'Q2',   '   12/11/19 17:55:46   ',  0,  1,
111,    58283,  '61',   '   12/11/19 17:55:57   ',  11, 1,
111,    58284,  'Q3',   '   12/11/19 17:55:57   ',  0,  1,
111,    58285,  '4',    '   12/11/19 17:56:13   ',  16, 1,
111,    58286,  'Q4',   '   12/11/19 17:56:13   ',  0,  1,
111,    58288,  '3',    '   12/11/19 17:56:42   ',  29, 1,
111,    58289,  'Interim partial',  '   12/11/19 17:56:42   ',  0,  1,
111,    58290,  'Q5',   '   12/11/19 17:56:42   ',  0,  1,
111,    58291,  '3',    '   12/11/19 17:56:59   ',  17, 1,
111,    58292,  'Q6',   '   12/11/19 17:56:59   ',  0,  1,
111,    58293,  '3',    '   12/11/19 17:57:25   ',  26, 1,
111,    58294,  'Q7',   '   12/11/19 17:57:25   ',  0,  1,
111,    58295,  '3',    '   12/11/19 17:57:38   ',  13, 1,
111,    58296,  'Q8',   '   12/11/19 17:57:38   ',  0,  1,
111,    58297,  '3',    '   12/11/19 17:57:50   ',  12, 1,
111,    58298,  'Q9',   '   12/11/19 17:57:50   ',  0,  1,
111,    58299,  'Timeout',  '   12/11/19 17:58:09   ',  19, 1,
111,    58300,  'Q9',   '   12/11/19 17:58:09   ',  0,  1,
111,    58301,  '10',   '   12/11/19 17:58:32   ',  23, 1,
111,    58302,  'Q10',  '   12/11/19 17:58:32   ',  0,  1,
111,    58303,  '1',    '   12/11/19 17:58:49   ',  17, 1,
111,    58304,  'Q11',  '   12/11/19 17:58:49   ',  0,  1,
111,    58307,  '3',    '   12/11/19 17:59:02   ',  13, 1,
111,    58308,  'Q12',  '   12/11/19 17:59:02   ',  0,  1,
111,    58309,  'Timeout',  '   12/11/19 17:59:23   ',  21, 1,
111,    58310,  'Q13',  '   12/11/19 17:59:23   ',  0,  1,
111,    58311,  'Timeout',  '   12/11/19 17:59:44   ',  21, 1,
111,    58312,  'Q13',  '   12/11/19 17:59:44   ',  0,  1,
111,    58313,  '4',    '   12/11/19 17:59:51   ',  7,  1,
111,    58314,  'Q14',  '   12/11/19 17:59:51   ',  0,  1,
111,    58318,  '2',    '   12/11/19 18:00:19   ',  28, 1,
111,    58319,  'Q15',  '   12/11/19 18:00:19   ',  0,  1,
111,    58320,  '3',    '   12/11/19 18:00:36   ',  17, 1,
111,    58321,  'Q16',  '   12/11/19 18:00:36   ',  0,  1,
111,    58322,  '4',    '   12/11/19 18:01:04   ',  28, 1,
111,    58323,  'Q17',  '   12/11/19 18:01:04   ',  0,  1,
111,    58324,  '1',    '   12/11/19 18:01:18   ',  14, 1,
111,    58325,  'Q18',  '   12/11/19 18:01:18   ',  0,  1,
111,    58327,  '4',    '   12/11/19 18:01:42   ',  24, 1,
111,    58328,  'Q19',  '   12/11/19 18:01:42   ',  0,  1,
111,    58329,  '1',    '   12/11/19 18:01:56   ',  14, 1,
111,    58330,  'Q20',  '   12/11/19 18:01:56   ',  0,  1,
111,    58331,  '1',    '   12/11/19 18:02:19   ',  23, 1,
111,    58332,  'Q21',  '   12/11/19 18:02:19   ',  0,  1,
111,    58333,  '1',    '   12/11/19 18:02:28   ',  9,  1,
111,    58334,  'Q22',  '   12/11/19 18:02:28   ',  0,  1,
111,    58335,  '2',    '   12/11/19 18:02:52   ',  24, 1,
111,    58336,  'Completed',    '   12/11/19 18:02:52   ',  0,  0,
111,    58337,  'Complete Message', '   12/11/19 18:02:52   ',  0,  0,
111,    58338,  'Thank you',    '   12/11/19 18:02:52   ',  0,  0),
nrow=87, ncol=6, byrow=T,
dimnames=list(c(NULL), c("Respondent.ID", "order.ID", "Response", "Datetime", "difftime", "Include"))))

我尝试添加指示开始和停止位置的列，但不知道如何索引开始和停止之间的行。

更新

我发现了一个案例，我们切断了调查，最终切断了一些受访者。发生这种情况时，他们不会在交互数据中获得停止代码。结果，rle 没有获取我们想要捕获的最终运行长度。我一直在摆弄它，但这对我来说是新事物，所以我想我会发布一个上面的测试数据集版本来模拟这种情况。

test2 <- data.frame(matrix(c(
111,    2340,   'Enqueueing call',  '12/2/19 14:53:57', NA,
111,    6174,   'Call expired, will be retried in next schedule window',    '12/2/19 20:14:22', 19225,
111,    10386,  'Answer',   '12/3/19 15:48:56', 70474,
111,    10387,  'Contacted',    '12/3/19 15:48:56', 0,
111,    10388,  'Intro',    '12/3/19 15:48:56', 0,
111,    10389,  'Timeout',  '12/3/19 15:49:16', 20,
111,    10390,  'Intro',    '12/3/19 15:49:16', 0,
111,    10391,  '1',    '12/3/19 15:49:30', 14,
111,    10392,  'Started',  '12/3/19 15:49:30', 0,
111,    10393,  'Q1',   '12/3/19 15:49:30', 0,
111,    10394,  '1',    '12/3/19 15:49:45', 15,
111,    10395,  'Q2',   '12/3/19 15:49:45', 0,
111,    10396,  'Timeout',  '12/3/19 15:49:54', 9,
111,    10397,  'Q2',   '12/3/19 15:49:54', 0,
111,    10398,  'Timeout',  '12/3/19 15:50:03', 9,
111,    10399,  'Q2',   '12/3/19 15:50:03', 0,
111,    10400,  'Timeout',  '12/3/19 15:50:11', 8,
111,    17658,  'Timeout. Call failed.',    '12/4/19 17:50:27', 93616,
111,    19932,  'Call expired, will be retried in next schedule window',    '12/4/19 20:45:17', 10490,
111,    25647,  'Call expired, will be retried in next schedule window',    '12/5/19 20:33:27', 85690,
111,    31516,  'Call expired, will be retried in next schedule window',    '12/6/19 20:17:18', 85431,
111,    36781,  'Call expired, will be retried in next schedule window',    '12/7/19 20:02:16', 85498,
111,    38718,  'Answer',   '12/8/19 10:24:07', 51711,
111,    38719,  'Q2',   '12/8/19 10:24:07', 0,
111,    38720,  'Timeout',  '12/8/19 10:24:16', 9,
111,    38721,  'Q2',   '12/8/19 10:24:16', 0,
111,    38722,  'Timeout',  '12/8/19 10:24:24', 8,
111,    38723,  'Q2',   '12/8/19 10:24:24', 0,
111,    38724,  'Timeout',  '12/8/19 10:24:33', 9,
111,    45339,  'Timeout. Call failed.',    '12/9/19 12:25:08', 93635,
111,    49026,  'Call expired, will be retried in next schedule window',    '12/9/19 20:31:26', 29178,
111,    53972,  'Call expired, will be retried in next schedule window',    '12/10/19 20:14:38',    85392,
111,    58277,  'Answer',   '12/11/19 17:55:29',    78051,
111,    58278,  'Q2',   '12/11/19 17:55:29',    0,
111,    58279,  'Timeout',  '12/11/19 17:55:38',    9,
111,    58280,  'Q2',   '12/11/19 17:55:38',    0,
111,    58281,  'Timeout',  '12/11/19 17:55:46',    8,
111,    58282,  'Q2',   '12/11/19 17:55:46',    0,
111,    58283,  '61',   '12/11/19 17:55:57',    11,
111,    58284,  'Q3',   '12/11/19 17:55:57',    0,
111,    58285,  '4',    '12/11/19 17:56:13',    16,
111,    58286,  'Q4',   '12/11/19 17:56:13',    0,
111,    58288,  '3',    '12/11/19 17:56:42',    29,
111,    58289,  'Interim partial',  '12/11/19 17:56:42',    0,
111,    58290,  'Q5',   '12/11/19 17:56:42',    0,
111,    58291,  '3',    '12/11/19 17:56:59',    17,
111,    58292,  'Q6',   '12/11/19 17:56:59',    0,
111,    58293,  '3',    '12/11/19 17:57:25',    26,
111,    58294,  'Q7',   '12/11/19 17:57:25',    0,
111,    58295,  '3',    '12/11/19 17:57:38',    13,
111,    58296,  'Q8',   '12/11/19 17:57:38',    0,
111,    58297,  '3',    '12/11/19 17:57:50',    12,
111,    58298,  'Q9',   '12/11/19 17:57:50',    0,
111,    58299,  'Timeout',  '12/11/19 17:58:09',    19,
111,    58300,  'Q9',   '12/11/19 17:58:09',    0,
111,    58301,  '10',   '12/11/19 17:58:32',    23,
111,    58302,  'Q10',  '12/11/19 17:58:32',    0,
111,    58303,  '1',    '12/11/19 17:58:49',    17,
111,    58304,  'Q11',  '12/11/19 17:58:49',    0,
111,    58307,  '3',    '12/11/19 17:59:02',    13,
111,    58308,  'Q12',  '12/11/19 17:59:02',    0,
111,    58309,  'Timeout',  '12/11/19 17:59:23',    21,
111,    58310,  'Q13',  '12/11/19 17:59:23',    0,
111,    58311,  'Timeout',  '12/11/19 17:59:44',    21,
111,    58312,  'Q13',  '12/11/19 17:59:44',    0,
111,    58313,  '4',    '12/11/19 17:59:51',    7,
111,    58314,  'Q14',  '12/11/19 17:59:51',    0,
111,    58318,  '2',    '12/11/19 18:00:19',    28
),
nrow=68, ncol=5, byrow=T,
dimnames=list(c(NULL), c("Respondent.ID", "order.ID", "Interaction", "Datetime", "difftime"))))

【问题讨论】：

标签： r indexing time

【解决方案1】：

这些类型的问题通常可以通过运行长度编码来解决。假设在下面的代码中，“test”是您发布的第一个数据集，“new”是第二个数据集。

我们首先需要弄清楚您数据中的哪些条目是开始和停止位置或“类别”。

start <- "Answer"
stop <- c("Timeout. Call failed.", "Completed")

cat <- ifelse(test$Interaction %in% start, "start",
              ifelse(test$Interaction %in% stop, "stop", "other"))

随后，我们可以运行长度编码cat，并获得开始和停止位置，并根据这些确定哪些行应该不同。

rle <- rle(cat)
start_positions <- which(rle$values == "start")
stop_positions  <- which(rle$values == "stop")

# Points that need to be changed are after starts and before stops
# The intersect ensures that we don't flag a 'start', 'other', 'start' sequence to be changed
change <- intersect(start_positions + 1, stop_positions - 1)

最后，我们需要将change编码为1，其余编码为0并放回数据中：

rle$values <- rep.int(0, length(rle$values))
rle$values[change] <- 1

test$Include <- inverse.rle(rle)

我们可以通过制作混淆矩阵来检查结果是否正确：

> table(test = test$Include, new = new$Include)
    new
test  0  1
   0 16  0
   1  0 71

编辑：如果没有适当的停止位置，以下应该可以工作。我假设数据是在Respondent.ID 上排序的，所以如果这个假设不正确，那么这可能不是最好的方法。我不清楚是否应该包括或排除无证停留。在下面的代码中，它们被包含在内。

start <- "Answer"
stop <- c("Timeout. Call failed.", "Completed")

cat <- ifelse(test2$Interaction %in% start, "start",
              ifelse(test2$Interaction %in% stop, "stop", "other"))

responder_rle <- rle(as.character(test2$Respondent.ID))
cat[cumsum(responder_rle$lengths)] <- "unplannedstop"

rle <- rle(cat)
start_positions <- which(rle$values == "start")
stop_positions  <- which(rle$values == "stop")
unplanned_stop  <- which(rle$values == "unplannedstop")

change <- c(intersect(start_positions + 1, 
                      c(stop_positions, unplanned_stop) - 1), 
            unplanned_stop)

rle$values <- rep.int(0, length(rle$values))
rle$values[change] <- 1

test2$Include <- inverse.rle(rle)

如果不应包含计划外停止，您可以将以下内容用于change 变量：

change <- intersect(start_positions + 1, c(stop_positions, unplanned_stop) - 1)

【讨论】：

嗨@teunbrand，这个解决方案在很大程度上有效，但我遇到了一种特殊情况，它没有。当我们自己关闭调查时，一些受访者被切断了，他们的响应数据中没有停止代码。在这些情况下，不会标记该运行长度的起始位置。我将设置一个示例数据集并发布它。
我已经发布了代码以在编辑中解决这种情况。我没有针对多个响应者的任何测试数据，因此请务必检查输出是否符合预期。
非常感谢！！
刚刚测试过，效果很好，再次感谢！