Question

我的任务需要按周进行多次api调用。我想通过purrr将所有这些数据与简明的代码一起追加，但我很难在我的函数中使Sys.sleep工作。

以下是一些虚假数据。我假装这是API后端：

#call 1
week1 <- as_tibble(list(week = rep(1,7),value = rnorm(7,0,1)))

#call 2
week2 <- as_tibble(list(week = rep(2,7),value = rnorm(7,0,1)))

#call 3
week3 <- as_tibble(list(week = rep(3,7),value = rnorm(7,0,1)))

#append
pretend_full_api_data <- bind_rows(week1, week2, week3)

此功能有效

by_api_call <- function(weeknumber){
  weeki <- pretend_full_api_data %>%
    filter(week == weeknumber)
}

seq(from = 1, to = 3) %>% 
map_df(~by_api_call(.))

问题是API通常需要休眠时间。所以我试着修改如下：

by_api_call <- function(weeknumber){
  weeki <- pretend_full_api_data %>%
    filter(week == weeknumber)
    Sys.sleep(2) #required sleep for API calls
}

seq(from = 1, to = 3) %>% 
map_df(~by_api_call(.))

上面的代码返回0 rows。上述功能有什么问题？

Answer 1

实际上这不起作用，因为函数by_api_call没有返回任何内容。您需要添加函数return以实际从函数中检索某些内容。

如果您需要查看技术讨论，请参阅Hadley Wickham撰写的Advanced R一书（R语言中最有价值的玩家）。

您的代码需要进行此调整，以实际执行所需的工作

# API stuff..

# this is the function that actually works...

by_api_call <- function(weeknumber){
  weeki <- pretend_full_api_data %>%
    filter(week == weeknumber)
    Sys.sleep(2) #required sleep for API calls

  return(weeki) # return is your friend...
}

# calling the API function
week_numbers <- seq(from = 1, to = 3, by = 1)

(map_df(week_numbers, by_api_call))

使用purrr将行附加到数据框

1 个答案: