修改

Question

考虑这个令牌数据集：

Data = structure(list(txs = c(-50, -750, -35, -5.96, -61.5, -42.07, 
-142.4, -500, 132, -154.89, -109.51, -2000, -50, -40, -24.98, 
-15.6, -50, -147.72, -20, -6.6, -5, -20, -13.48, -7.25, -54.09, 
-200, -124.11, -30, -50, -30, 400, -10, -0.95, -4.1, -10000, 
30, -1.99, 74.03, -6.95, -2.96, -29, -403.6, -6, -6, 5250, -513.57, 
-300, -10, -500, -20, -6.45, -7.26, -40, -50, -13.14, 321.29, 
-18, 100, -5.5, -25, -59.2, -10.75, -3.2, 270, 65.8, -11.6, -104.78, 
-99.39, 0.1, -50, -80, -50, -371.44, -78, 270, -6.3, 40, -2.5, 
-29.99, -189.48, -400, -0.29, -20, -6.55, -987.37, -1400, -0.49, 
-20, -29.04, -65, -40, -27.5, -17.37, -10, -1092.84, -5.5, -69.93, 
-15.07, -400, -4.8), week = structure(c(1439157600, 1454281200, 
1471212000, 1445205600, 1448233200, 1451862000, 1449442800, 1453676400, 
1460325600, 1460930400, 1445205600, 1454281200, 1460930400, 1444600800, 
1462140000, 1471816800, 1443996000, 1448838000, 1479682800, 1453071600, 
1447023600, 1473631200, 1465768800, 1433109600, 1445205600, 1433714400, 
1466978400, 1441576800, 1459116000, 1451862000, 1436133600, 1440367200, 
1456095600, 1458514800, 1456700400, 1450047600, 1440972000, 1446418800, 
1465164000, 1441576800, 1442181600, 1453071600, 1461535200, 1460930400, 
1438552800, 1464559200, 1447628400, 1434924000, 1437343200, 1436738400, 
1443391200, 1438552800, 1440972000, 1446418800, 1446418800, 1453071600, 
1453071600, 1457305200, 1444600800, 1462140000, 1435528800, 1457305200, 
1437948000, 1440972000, 1437948000, 1433109600, 1461535200, 1453676400, 
1454886000, 1454281200, 1441576800, 1441576800, 1471212000, 1453071600, 
1451862000, 1442786400, 1443391200, 1439762400, 1436133600, 1461535200, 
1442181600, 1468188000, 1442181600, 1453676400, 1466373600, 1443391200, 
1450652400, 1454886000, 1439157600, 1441576800, 1463954400, 1442181600, 
1446418800, 1454886000, 1476050400, 1461535200, 1456700400, 1456700400, 
1435528800, 1456700400), class = c("POSIXct", "POSIXt"), tzone = ""), 
    num_c = c(1219, 1257, 1195, 33, 1105, 1223, 1257, 1317, 486, 
    1227, 477, 1039, 1238, 1008, 1137, 1294, 1070, 596, 1295, 
    1354, 1010, 1294, 1348, 1254, 19, 1185, 24, 1287, 1198, 955, 
    1324, 1293, 1343, 1162, 1272, 972, 972, 179, 1343, 1105, 
    1085, 1020, 947, 1375, 1005, 477, 596, 1198, 928, 1137, 1263, 
    1237, 1054, 1288, 1185, 1115, 1257, 1301, 1294, 1185, 1039, 
    957, 1131, 33, 477, 1258, 477, 1039, 1362, 1246, 596, 1010, 
    972, 1238, 477, 1296, 972, 1148, 1105, 24, 553, 1297, 1288, 
    1223, 789, 1298, 1082, 1353, 1030, 1287, 1203, 1008, 1294, 
    1227, 1298, 1203, 1346, 1010, 19, 1303)), .Names = c("txs", 
"week", "num_c"), row.names = c(NA, -100L), class = "data.frame")

它有三列：

一个名为num_c：这是一个客户编号
一个名为week的人：这是周的星期一的日期 num_c下了订单。
一个名为'txs'：每周帐户余额和客户（num_c）

现在，对于每个客户端，我都想使用dplyr来扩展（加入）这个数据集这样我就可以在每个星期的范围内获得一行客户已下订单。

此外， new 空单元格（对于txs）应填充NA。对应于原始（未扩展）数据集中存在的那些周/客户端的txs值应保留其原始值;

我试过了：

 library(dplyr)
 stretch_Data = Data %>%
                group_by(num_c) %>%
                right_join(seq(min(week), max(week), by = 'week'), by = "week")

但我得到了;

Error in seq(min(week), max(week), by = "week") : object 'week' not found

这很愚蠢，因为数据确实包含一周的列（非常感谢你）。

我做错了什么？

修改

感谢@ mt1022的解决方案（下面的评论）。这很安静聪明。但仍然存在一个问题：是否可以包含数据中的所有列？ - 为了表达我的观点，我添加了一个：txs--那么缺失的值（那些对应于没有购买的周数的那些）应该由NA填充（就像在联合中一样）。与我们购买的客户*周日期相对应的单元格应保留其原始值。

本质上在new（扩展表）上执行na.omit（）应返回原始表;就像在关节上一样;

Answer 1

使用@ mt1002中的解决方案，只需一个简单的添加即可获得txs列。我已经sum(txs)获得了一个值，根据您的需要，这也可以是min(txs)或max(txs)的单个值。

Data %>% group_by(num_c) %>% 
  summarise(week = list(seq(min(week), max(week), by = 'week')),
            txs = sum(txs)) %>% 
  unnest(week)

澄清之后，这是我提出的解决方案，除了每周每位用户的订单数量之外，没有订单的数周有NA值。您也可以使用左连接（包括上述查询到df），按num_c的订单加入周列表。

library(lubridate)
a <- data.frame(week = rep(seq(1,52,1)))
Data %>% 
  group_by(num_c) %>% 
  mutate(week_num = week(week)) %>% 
  group_by(num_c, week_num) %>% 
  summarise(txs = sum(txs),
            number_orders = n()) %>% 
  full_join(a, by = c("week_num"="week")) %>% 
  ungroup() %>% 
  arrange(week_num)

按组ID在日期范围内左连接

修改

1 个答案: