r - rmongodb $或查询构造

r - rmongodb $or query construction

背景

我有 GTFS 数据存储在本地 mongodb 数据库中。

calendartable长得像

field      | type
service_id | varchar
monday     | int (0 or 1)
tuesday    | int (0 or 1)
...
sunday     | int (0 or 1)

任务

我想 select 任何工作日(周一到周五)= 1 的所有 service_id,使用 r 中的 rmongodb 包。

在 SQL 中类似于:SELECT service_id FROM calendar WHERE monday = 1 OR tuesday = 1 OR ... OR friday = 1

详情

使用 Robomongo GUI 时,查询是:

db.getCollection('calendar').find({"$or" : 
    [{'monday':1},
    {'tuesday':1},
    {'wednesday':1},
    {'thursday':1},
    {'friday':1}]
})

其中 returns 8 个文档(见图)

因此,在 r 中,我试图构建相同的 or 查询,将 return 得到相同的结果,但我运气不好。

library(rmongodb)
library(jsonlite)
## connect to db
mongo <- mongo.create()
mongo.is.connected(mongo)
db <- "temp"

## days for which I want a service:
serviceDays <- c("monday","tuesday","wednesday","thursday","friday")

尝试 0:

## create list as the 'query' condition
ls <- list("$or" = 
         list("monday" = 1L, 
              "tuesday" = 1L, 
              "wednesday" = 1L, 
              "thursday" = 1L, 
              "friday" = 1L))

services <- mongo.find.all(mongo, "temp.calendar", query=ls)
## returns error:
Error in mongo.find(mongo, ns, query = query, sort = sort, fields = fields,  : 
  find failed with unknown error.

尝试 1:

## paste the string together
js <- paste0('{"', serviceDays, '":[',1L,']}', collapse=",")
js <- paste0('{"$or" :[', js, ']}')
## this string has been validated at jsonlint.com

bs <- mongo.bson.from.JSON(js)
## run query
services <- mongo.find.all(mongo, "temp.calendar", query=bs)
## result
> services
list()    ## empty list

## manually writing the JSON string doesn't work either
# js <- '{"$or" : [{"monday":[1]},{"tuesday":[1]},{"wednesday":[1]},{"thursday":[1]},{"friday":[1]}]}'

尝试 2:

## create the or condition using R code
l <- as.list(sapply(serviceDays, function(y) 1L))
bs <- mongo.bson.from.list(list("$or" = list(l)))
## run query
services <- mongo.find.all(mongo, "temp.calendar", query=bs)
## result
> length(services)
[1] 2    ## 2 documents returned

两个文档 returned 是针对 service_ids 的,其中所有星期一、星期二、星期三、星期四、星期五 = 1。即,它似乎使用了 AND子句,而不是 OR.

尝试 3:

## deconstruct the JSON string (attempt 1)
js <- fromJSON(js, simplifyVector=FALSE)
bs <- mongo.bson.from.list(js)

## run query
services <- mongo.find.all(mongo, "temp.calendar", query=bs)
## result
> services
list()    ## empty list

我在 R 中的查询尝试有什么问题导致我无法获得与使用 Robomongo GUI 时相同的 8 个文档?

我的 'attempt 0' 很接近,但我缺少更多 list 个参数。

ls <- list("$or" = list(list("monday" = 1L), 
                    list("tuesday" = 1L), 
                    list("wednesday" = 1L), 
                    list("thursday"= 1L), 
                    list("friday" = 1L)))
## json string:
> toJSON(ls)
{"$or":[{"monday":[1]},{"tuesday":[1]},{"wednesday":[1]},{"thursday":[1]},{"friday":[1]}]} 
## run query:
services <- mongo.find.all(mongo, "temp.calendar", query=ls)

## result
length(services)
[1] 8