简体   繁体   English

R中的ODBC / DBI不会写入R中具有非默认架构的表

[英]ODBC/DBI in R will not write to a table with a non-default schema in R

The Issue 问题

When trying to write to a table with a non-default schema, dbWriteTable in the package DBI, writes to default.non-default.tablename rather than writing to non-default.tablename . 尝试写入具有非默认架构的表时,DBI包中的dbWriteTable写入default.non-default.tablename而不是写入non-default.tablename I know that non-default.tablename exists because it's showing up in my SSMS database. 我知道non-default.tablename存在,因为它出现在我的SSMS数据库中。

Reproducible Example/What I've Tried 可重复的例子/我试过的

Create this table in SQL Server with a non-default schema 'guest'. 使用非默认架构“guest”在SQL Server中创建此表。 I am placing it in a database called 'SAM': 我将它放在一个名为'SAM'的数据库中:

CREATE TABLE guest.MikeTestTable(
[a] [float] NULL,
[b] [float] NULL,
[c] [varchar](255) NULL)

#Create a df to insert into guest.MikeTestTable
df <- data.frame(a = c(10, 20, 30),
                 b = c(20, 40, 60),
                 c = c("oneT", "twoT", "threeT"))

#Create a connection:
con <- DBI::dbConnect(odbc::odbc(),
             .connection_string = "Driver={SQL Server};
                                   server=localhost;
                                   database=SAM;
                                   trustedConnection=true;")

#Try to write contents of df to the table using `dbWriteTable`
DBI::dbWriteTable(conn = con,
                  name = "guest.MikeTestTable",
                  value = df,
                  append = TRUE)

#Create a query to read the data from `"guest.MikeTestTable"`:
q <- "SELECT [a]
  ,[b]
  ,[c]
  FROM guest.MikeTestTable"

##Read the table into R to show that nothing actually got written to the 
##table but that it recognizes `guest.MikeTestTable` does exist:
DBI::dbGetQuery(con, q)

[1] a b c
<0 rows> (or 0-length row.names)

I thought this was a weird result so I opened up my SSMS and lo and behold, the table dbo.guest.MikeTestTable was created. 我认为这是一个奇怪的结果所以我打开了我的SSMS并且看到了dbo.guest.MikeTestTable表的创建。 Any help would be much appreciated. 任何帮助将非常感激。

The CRAN release last week (related to the issue @user111417 linked to) resolves this using the new DBI::Id() function, where the schema and table names are separate and explicit. 上周的CRAN版本(与@ user111417链接的问题相关)使用新的DBI::Id()函数解决了这个问题,其中模式和表名是独立且明确的。 Here's an example. 这是一个例子。

library(magrittr)
table_id <- DBI::Id(
  schema  = "schema_1",
  table   = "car"
)

ds <- mtcars %>% 
  tibble::rownames_to_column("car")

# Create the Table
channel <- DBI::dbConnect(
  drv   = odbc::odbc(),
  dsn   = "cdw_cache"
)
result <- DBI::dbWriteTable(
  conn        = channel,
  name        = table_id,
  value       = ds,
  overwrite   = T,
  append      = F
)

DBI::dbGetQuery(channel, "SELECT COUNT(*) FROM schema_1.car")
# Produces `1 32`

DBI::dbExistsTable(channel, table_id)
# Produces: [1] TRUE

DBI::dbDisconnect(channel)

(Thanks to Daniel Wood for help in https://github.com/r-dbi/odbc/issues/191 .) (感谢Daniel Wood在https://github.com/r-dbi/odbc/issues/191上的帮助。)

请看这里的答案

声明:本站的技术帖子网页,遵循CC BY-SA 4.0协议,如果您需要转载,请注明本站网址或者原文地址。任何问题请咨询:yoyou2525@163.com.

 
粤ICP备18138465号  © 2020-2024 STACKOOM.COM