如何根據第一個數據框中的數據對從第二個數據框提取行？

Question

我有兩個數據幀，df1（2000行）和df2（100,000行）。我想：

• 讀經/緯度組合從DF1
• 如同緯度/長從DF2
• 閱讀DF2的相應的 「M」 值
• 將它添加到DF1

我該怎麼辦？

df1 
lat  long  val 
33.29083 -109.19556 12.5000 
32.96583 -109.30972 11.5000 
33.04944 -109.29528 1.7500 
33.06444 -109.44167 1.7500 
33.61944 -110.92083 0.2500 
33.98000 -111.30278 3.7500 
33.79806 -110.49917 1.00008 
38.37972 -119.44917 3.2500 

而且

df2<-structure(list(lat = c(33.29, 32.96, 48.15, 48.1, 48.18, 48.14 
    ), lon = c(-109.19, -109.31, -124.69, -124.69, -124.68, -124.68 
    ), m = c(0.4713, 0.8998, 0.4891, 0.8418, 0.7998, 0.5292), flagar = c(0L, 
    1L, 0L, 1L, 1L, 0L), flagk = c(0L, 0L, 0L, 0L, 0L, 0L), flagsi = c(0L, 
    0L, 0L, 0L, 0L, 0L), flags2o = c(0L, 0L, 0L, 0L, 0L, 0L), flagap = c(0L, 
    0L, 0L, 0L, 0L, 0L), flagmt = c(0L, 0L, 0L, 0L, 0L, 0L), flagcn = c(0L, 
    0L, 0L, 0L, 0L, 0L), flagkf = c(1L, 1L, 1L, 1L, 1L, 1L), flagrd = c(1L, 
    1L, 1L, 1L, 1L, 1L), flagrv = c(2L, 2L, 2L, 2L, 2L, 2L), flagpt = c(1L, 
    1L, 1L, 1L, 1L, 1L)), .Names = c("lat", "lon", "m", "flagar", 
    "flagk", "flagsi", "flags2o", "flagap", "flagmt", "flagcn", "flagkf", 
    "flagrd", "flagrv", "flagpt"), row.names = c(NA, 6L), class = "data.frame") 

Answer 1

使用data.table包，你可以做到這一點，如下所示：

library(data.table) 
# convert the dataframes to datatables (which are an extended form of a dataframe) 
# and round the lat & lon variables to one decimal 
setDT(df1)[, `:=` (lat = round(lat,1), lon = round(lon,1))] 
setDT(df2)[, `:=` (lat = round(lat,1), lon = round(lon,1))] 

# joining the m column of df2 to df1 based on exact matches of the rounded lat/lon values 
# m = i.m transfers the matching values to the df1 dataframe 
df1[df2, m := i.m, on=c("lat","lon")] 

這給：

> df1 
    lat lon  val  m 
1: 33.3 -109.2 12.50000 0.4713 
2: 33.0 -109.3 11.50000 0.8998 
3: 33.0 -109.3 1.75000 0.8998 
4: 33.1 -109.4 1.75000  NA 
5: 33.6 -110.9 0.25000  NA 
6: 34.0 -111.3 3.75000  NA 
7: 33.8 -110.5 1.00008  NA 
8: 38.4 -119.4 3.25000  NA 

如果你想匹配最短距離，見this answer爲例。