Queries in bigrquery being billed for much more data than if run in BigQuery SQL editor

[sharonm]

Update: I was told on Twitter that's due to a minimum for data queries. I still don't understand why in the Web interface query history the bigrquery queries show those minimums but the ones I run in the SQL BigQuery Web editor don't, but I guess that must be it.

If I run this SQL in the BigQuery Web-based SQL editor

SELECT homeTeamName from bigquery-public-data.baseball.schedules

it says it will use 21.1 KiB. But if I run the query in R and bigrquery

library(bigrquery)
library(dplyr)
library(DBI)
data_project_id <- "bigquery-public-data"
billing_project_id <- "my_project_id"
database_id <- "baseball"

con <- dbConnect(
  bigrquery::bigquery(),
  project = data_project_id,
  billing = billing_project_id,
  dataset = database_id
)

sql <- "SELECT homeTeamName from bigquery-public-data.baseball.schedules"
my_results <- dbGetQuery(con, sql)

Results say 10.49 MB were billed.

Anyone know why the same query in bigrquery uses so much more analysis data? Thanks.

[nirgrahamuk]

@sharonm

it says it will use 21.1 KiB

if this is a prediction, perhaps its inaccurate ?