Does duckdb allow Concurrent read?

[shahronak47]

I am having a duckdb file which has a table (master_file) and I want to know if concurrent read is allowed or not. Based on reading several posts online https://duckdb.org/docs/connect/concurrency and chatting with chatGPT the answer seems to be yes and I want to believe it but I am struggling to prove this.

I have tried this chunk of code for the same

library(DBI)
library(duckdb)
library(future.apply)

read_query <- function(i) {
  con <- dbConnect(duckdb::duckdb(), dbdir = "master_file.duckdb")
  Sys.sleep(2) # Deliberately adding a delay
  start_time <- Sys.time()
  res <- dbGetQuery(con, "select * from master_file")
  end_time <- Sys.time()
  dbDisconnect(con)
  return(data.frame(
    query_id = i,
    start_time = start_time,
    end_time = end_time,
    duration = as.numeric(difftime(end_time, start_time, units = "secs"))
  ))
}

# Run multiple SELECT queries in parallel
plan(multisession, workers = 4)
select_results <- future_lapply(1:4, read_query)

However, this returns an error

Error in .local(drv, ...) :
rapi_startup: Failed to open database: {"exception_type":"IO","exception_message":"File is already open in \nC:\PROGRA1\R\R-441.0\bin\x64\Rscript.exe (PID 1012)"}

My idea was that if there is an overlap between start and end time between two processes then we can say that multiple reads was allowed but this didn't work. How do I prove that multiple read in duckdb is allowed?

[Saarialho]

I believe it is possible, try dbConnect(duckdb::duckdb(), dbdir = "master_file.duckdb", read_only = TRUE) but using read_only = TRUE you cannot write to duckdb from multiple processes