I really thought I had a solution to this problem when I learned about the additional parameter 'sort' that was defaulted to TRUE in merge(). However, setting this to false did not help. Below is a demo of my code, with the results I am getting and the results I want:
df2 = structure(list(player = c("Marvin Williams", "Spencer Hawes",
"Jeremy Lin", "Kemba Walker", "P.J. Hairston", "Rudy Gay", "Rajon Rondo",
"DeMarcus Cousins", "Ben McLemore", "Willie Cauley-Stein"), global.player.id = c(263884L,
329824L, 340730L, 462980L, 609567L, 266358L, 262882L, 509450L,
604898L, 699950L), team.name = c("Hornets", "Hornets", "Hornets",
"Hornets", "Grizzlies", "Kings", "Kings", "Kings", "Kings", "Kings"
)), .Names = c("player", "global.player.id", "team.name"), class = "data.frame", row.names = c(47L,
48L, 52L, 53L, 225L, 389L, 390L, 395L, 398L, 401L))
df1 = structure(list(global.player.id = c(-1L, 262882L, 266358L, 509450L,
604898L, 699950L, 263884L, 329824L, 340730L, 462980L, 609567L,
-1L, 262882L, 266358L, 509450L, 604898L, 699950L, 263884L, 329824L,
340730L, 462980L, 609567L, -1L, 262882L, 266358L), x_loc = c(47.17753,
13.57165, 46.45843, 26.68803, 52.16717, 47.20201, 60.097, 47.20201,
52.16717, 65.1302, 46.45843, 47.19141, 13.61702, 46.5355, 26.71856,
52.25433, 47.27324, 60.08215, 47.27324, 52.25433, 65.11267, 46.5355,
46.82163, 13.66478, 46.57545), y_loc = c(26.44326, 25.18298,
18.46573, 25.48557, 33.09177, 31.09372, 22.79717, 31.09372, 33.09177,
26.39671, 18.46573, 26.5187, 25.17431, 18.42014, 25.53807, 33.11185,
31.01197, 22.76307, 31.01197, 33.11185, 26.40227, 18.42014, 26.72834,
25.17784, 18.35961), order = c(1, 2, 3, 4, 5, 6, 7, 8, 9, 10,
11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25)), .Names = c("global.player.id",
"x_loc", "y_loc", "order"), row.names = c("1", "2", "3", "4",
"5", "6", "7", "8", "9", "10", "11", "12", "13", "14", "15",
"16", "17", "18", "19", "20", "21", "22", "23", "24", "25"), class = "data.frame")
The above are the dataframes I am working with. I want to keep the order of df1 when i merge df2 onto it. I am dealing with time series data here, so the order of the dataframe is important. The order column in df1 is just to test whether df1 is being shuffled or not (I do not want to use extra code to sort on order after the merge).
Here is what I've tried:
merge(df1, df2, by = 'global.player.id', all.x = TRUE)
global.player.id x_loc y_loc order player team.name
1 -1 47.17753 26.44326 1 <NA> <NA>
2 -1 46.82163 26.72834 23 <NA> <NA>
3 -1 47.19141 26.51870 12 <NA> <NA>
4 262882 13.57165 25.18298 2 Rajon Rondo Kings
5 262882 13.61702 25.17431 13 Rajon Rondo Kings
6 262882 13.66478 25.17784 24 Rajon Rondo Kings
7 263884 60.08215 22.76307 18 Marvin Williams Hornets
8 263884 60.09700 22.79717 7 Marvin Williams Hornets
9 266358 46.53550 18.42014 14 Rudy Gay Kings
10 266358 46.45843 18.46573 3 Rudy Gay Kings
11 266358 46.57545 18.35961 25 Rudy Gay Kings
12 329824 47.27324 31.01197 19 Spencer Hawes Hornets
13 329824 47.20201 31.09372 8 Spencer Hawes Hornets
14 340730 52.16717 33.09177 9 Jeremy Lin Hornets
15 340730 52.25433 33.11185 20 Jeremy Lin Hornets
16 462980 65.13020 26.39671 10 Kemba Walker Hornets
17 462980 65.11267 26.40227 21 Kemba Walker Hornets
18 509450 26.71856 25.53807 15 DeMarcus Cousins Kings
19 509450 26.68803 25.48557 4 DeMarcus Cousins Kings
20 604898 52.16717 33.09177 5 Ben McLemore Kings
21 604898 52.25433 33.11185 16 Ben McLemore Kings
22 609567 46.53550 18.42014 22 P.J. Hairston Grizzlies
23 609567 46.45843 18.46573 11 P.J. Hairston Grizzlies
24 699950 47.20201 31.09372 6 Willie Cauley-Stein Kings
25 699950 47.27324 31.01197 17 Willie Cauley-Stein Kings
Originally in df1, order was sorted 1-25, and now it is all out of order. Clearly df1 was shuffled in a way I didn't want it to be. Here's the output when I pass sort = FALSE to the merge function:
merge(df1, df2, by = 'global.player.id', all.x = TRUE, sort = FALSE)
global.player.id x_loc y_loc order player team.name
1 262882 13.57165 25.18298 2 Rajon Rondo Kings
2 262882 13.61702 25.17431 13 Rajon Rondo Kings
3 262882 13.66478 25.17784 24 Rajon Rondo Kings
4 266358 46.53550 18.42014 14 Rudy Gay Kings
5 266358 46.45843 18.46573 3 Rudy Gay Kings
6 266358 46.57545 18.35961 25 Rudy Gay Kings
7 509450 26.71856 25.53807 15 DeMarcus Cousins Kings
8 509450 26.68803 25.48557 4 DeMarcus Cousins Kings
9 604898 52.16717 33.09177 5 Ben McLemore Kings
10 604898 52.25433 33.11185 16 Ben McLemore Kings
11 699950 47.20201 31.09372 6 Willie Cauley-Stein Kings
12 699950 47.27324 31.01197 17 Willie Cauley-Stein Kings
13 263884 60.08215 22.76307 18 Marvin Williams Hornets
14 263884 60.09700 22.79717 7 Marvin Williams Hornets
15 329824 47.27324 31.01197 19 Spencer Hawes Hornets
16 329824 47.20201 31.09372 8 Spencer Hawes Hornets
17 340730 52.16717 33.09177 9 Jeremy Lin Hornets
18 340730 52.25433 33.11185 20 Jeremy Lin Hornets
19 462980 65.13020 26.39671 10 Kemba Walker Hornets
20 462980 65.11267 26.40227 21 Kemba Walker Hornets
21 609567 46.53550 18.42014 22 P.J. Hairston Grizzlies
22 609567 46.45843 18.46573 11 P.J. Hairston Grizzlies
23 -1 47.17753 26.44326 1 <NA> <NA>
24 -1 46.82163 26.72834 23 <NA> <NA>
25 -1 47.19141 26.51870 12 <NA> <NA>
Also not what I wanted, as order is all out of order again.
Is there anyway to call the merge function without completely shuffling the first dataframe parameter passed, or am I completely out of luck. This seems like a major downfall of the merge() function if so. Thanks!
