fixes https://github.com/hannesmuehleisen/MonetDBLite/issues/19

[ajdamico]

No description provided.

[ajdamico]

@hannesmuehleisen i doubt you like the SELECT * here but i'm not finding the alternative. using something like

dbGetQuery(db, paste0("select columns.name , columns.type from sys.columns join sys.tables on \
	columns.table_id=tables.id where tables.name='", qname , "'"))	

returns monetdb sql types rather than the five possible dbDataType results:

"BOOLEAN","INTEGER","DOUBLE PRECISION","BLOB","STRING"

this pull request checks three things during an append..

1. number of columns matches, then
2. names of columns match, then
3. dbDataType of columns match OR data.frame has INTEGER appending into a DOUBLE PRECISION column already in the database

..and then fixes this coercion issue:

> library(DBI)
> db <- dbConnect(MonetDBLite::MonetDBLite())
> dbWriteTable(db,'mtcars',mtcars)
> x<-mtcars
> x$mpg<-as.integer(x$mpg)
> dbWriteTable(db,'mtcars',x,append=T)
Error in .local(conn, name, value, ...) : 
  Failed to insert data: MALException:embedded:Could not convert column 0 mpg to type 10 

thanks

[ajdamico]

does https://github.com/hannesmuehleisen/MonetDBLite/blob/Jul2017Lite-R/tests/testthat/test_02_dbi.R#L330 require that i revise this PR? thanks