问题
This is a follow-up question from Extract multiple data.frames from one with selection criteria.
Let's say the data is the same as in above example
df <- data.frame(x1 = runif(1000), x2 = runif(1000), x3 = runif(1000),
split = sample( c('SPLITMEHERE', 'OBS'), 1000, replace=TRUE, prob=c(0.1, 0.9) ))
Basically, I need more general solution than the one in the quoted example.
Namely, some counties in some months (every month is a .txt file) have only 1 table, and therefore only one 'SPLITMEHERE'.
For example, fourth county consists of only one table, which means this county's group should end at the first SPLITMEHERE, and not the second. Also, the last county consists of three tables, but this doesn't matter as it is at the end so I can easily merge the last group.
The problem is, it's not always the fourth county the one which consists of one table, sometimes there are others.
Let's say we have this from the above dataset:
x1 x2 x3 split
1 0.591940061 0.635445182 0.304498259 SPLITMEHERE
2 0.510158838 0.170956885 0.881018211 OBS
3 0.938369076 0.495642515 0.171227120 OBS
4 0.366153042 0.464698494 0.550931566 OBS
5 0.051998873 0.222881187 0.934175135 OBS
6 0.706940809 0.735885367 0.666272118 SPLITMEHERE
7 0.244219533 0.340480033 0.144009797 OBS
8 0.546891246 0.024010211 0.151338479 OBS
9 0.032659978 0.174774606 0.576820824 OBS
10 0.641988559 0.575596526 0.911188682 OBS
11 0.111024861 0.969227957 0.643551420 OBS
12 0.179469011 0.052698538 0.199299193 OBS
13 0.199203707 0.429210222 0.525920379 SPLITMEHERE
14 0.837223042 0.556442838 0.881305105 OBS
15 0.628854814 0.874139058 0.199226364 OBS
16 0.618989684 0.784011205 0.038021599 OBS
17 0.421893407 0.394786134 0.519100402 OBS
18 0.126453054 0.926114653 0.687669218 OBS
19 0.739393898 0.938428464 0.110824400 OBS
20 0.582882966 0.198520021 0.942501112 OBS
21 0.143852453 0.963329219 0.993098109 OBS
22 0.249366828 0.242881240 0.486960755 OBS
23 0.060602695 0.797436479 0.432171847 SPLITMEHERE
24 0.013947914 0.028245990 0.489656647 OBS
25 0.795170730 0.541771474 0.122952446 OBS
26 0.786673408 0.284252650 0.305914856 OBS
27 0.591369056 0.321041728 0.285482027 OBS
28 0.899577535 0.468031873 0.588038383 SPLITMEHERE
29 0.955853329 0.552076328 0.825239050 OBS
30 0.634738808 0.050917396 0.730090024 OBS
Let's say there are three counties in the printed output, and I want three data frames as follows:
df1
1 0.510158838 0.170956885 0.881018211 OBS
2 0.938369076 0.495642515 0.171227120 OBS
3 0.366153042 0.464698494 0.550931566 OBS
4 0.051998873 0.222881187 0.934175135 OBS
5 0.244219533 0.340480033 0.144009797 OBS
6 0.546891246 0.024010211 0.151338479 OBS
7 0.032659978 0.174774606 0.576820824 OBS
8 0.641988559 0.575596526 0.911188682 OBS
9 0.111024861 0.969227957 0.643551420 OBS
10 0.179469011 0.052698538 0.199299193 OBS
df2
1 0.837223042 0.556442838 0.881305105 OBS
2 0.628854814 0.874139058 0.199226364 OBS
3 0.618989684 0.784011205 0.038021599 OBS
4 0.421893407 0.394786134 0.519100402 OBS
5 0.126453054 0.926114653 0.687669218 OBS
6 0.739393898 0.938428464 0.110824400 OBS
7 0.582882966 0.198520021 0.942501112 OBS
8 0.143852453 0.963329219 0.993098109 OBS
9 0.249366828 0.242881240 0.486960755 OBS
df3
1 0.013947914 0.028245990 0.489656647 OBS
2 0.795170730 0.541771474 0.122952446 OBS
3 0.786673408 0.284252650 0.305914856 OBS
4 0.591369056 0.321041728 0.285482027 OBS
5 0.955853329 0.552076328 0.825239050 OBS
6 0.634738808 0.050917396 0.730090024 OBS
Any ideas?
来源：https://stackoverflow.com/questions/44021913/multiple-data-frames-from-one-with-almost-random-selection-criteria