R: Add support for the `type`/`dtypes` in `duckdb_read_csv`?

I am trying to read the Fannie Mae data which is a darling of the larger than RAM world.
But I can't seem to read it into R. I checked the source code and it doesn't support types yet.
For example, in disk.frame, we can read them like this by specifying the types but duckdb can't at the moment.
library(disk.frame) 
setup_disk.frame() # limit to two worker so as not to blow the memory usage

#library(future)

#future::plan(sequential, split = TRUE)

lppub_column_names <- c("POOL_ID", "LOAN_ID", "ACT_PERIOD", "CHANNEL", "SELLER", "SERVICER",
                        "MASTER_SERVICER", "ORIG_RATE", "CURR_RATE", "ORIG_UPB", "ISSUANCE_UPB",
                        "CURRENT_UPB", "ORIG_TERM", "ORIG_DATE", "FIRST_PAY", "LOAN_AGE",
                        "REM_MONTHS", "ADJ_REM_MONTHS", "MATR_DT", "OLTV", "OCLTV",
                        "NUM_BO", "DTI", "CSCORE_B", "CSCORE_C", "FIRST_FLAG", "PURPOSE",
                        "PROP", "NO_UNITS", "OCC_STAT", "STATE", "MSA", "ZIP", "MI_PCT",
                        "PRODUCT", "PPMT_FLG", "IO", "FIRST_PAY_IO", "MNTHS_TO_AMTZ_IO",
                        "DLQ_STATUS", "PMT_HISTORY", "MOD_FLAG", "MI_CANCEL_FLAG", "Zero_Bal_Code",
                        "ZB_DTE", "LAST_UPB", "RPRCH_DTE", "CURR_SCHD_PRNCPL", "TOT_SCHD_PRNCPL",
                        "UNSCHD_PRNCPL_CURR", "LAST_PAID_INSTALLMENT_DATE", "FORECLOSURE_DATE",
                        "DISPOSITION_DATE", "FORECLOSURE_COSTS", "PROPERTY_PRESERVATION_AND_REPAIR_COSTS",
                        "ASSET_RECOVERY_COSTS", "MISCELLANEOUS_HOLDING_EXPENSES_AND_CREDITS",
                        "ASSOCIATED_TAXES_FOR_HOLDING_PROPERTY", "NET_SALES_PROCEEDS",
                        "CREDIT_ENHANCEMENT_PROCEEDS", "REPURCHASES_MAKE_WHOLE_PROCEEDS",
                        "OTHER_FORECLOSURE_PROCEEDS", "NON_INTEREST_BEARING_UPB", "PRINCIPAL_FORGIVENESS_AMOUNT",
                        "ORIGINAL_LIST_START_DATE", "ORIGINAL_LIST_PRICE", "CURRENT_LIST_START_DATE",
                        "CURRENT_LIST_PRICE", "ISSUE_SCOREB", "ISSUE_SCOREC", "CURR_SCOREB",
                        "CURR_SCOREC", "MI_TYPE", "SERV_IND", "CURRENT_PERIOD_MODIFICATION_LOSS_AMOUNT",
                        "CUMULATIVE_MODIFICATION_LOSS_AMOUNT", "CURRENT_PERIOD_CREDIT_EVENT_NET_GAIN_OR_LOSS",
                        "CUMULATIVE_CREDIT_EVENT_NET_GAIN_OR_LOSS", "HOMEREADY_PROGRAM_INDICATOR",
                        "FORECLOSURE_PRINCIPAL_WRITE_OFF_AMOUNT", "RELOCATION_MORTGAGE_INDICATOR",
                        "ZERO_BALANCE_CODE_CHANGE_DATE", "LOAN_HOLDBACK_INDICATOR", "LOAN_HOLDBACK_EFFECTIVE_DATE",
                        "DELINQUENT_ACCRUED_INTEREST", "PROPERTY_INSPECTION_WAIVER_INDICATOR",
                        "HIGH_BALANCE_LOAN_INDICATOR", "ARM_5_YR_INDICATOR", "ARM_PRODUCT_TYPE",
                        "MONTHS_UNTIL_FIRST_PAYMENT_RESET", "MONTHS_BETWEEN_SUBSEQUENT_PAYMENT_RESET",
                        "INTEREST_RATE_CHANGE_DATE", "PAYMENT_CHANGE_DATE", "ARM_INDEX",
                        "ARM_CAP_STRUCTURE", "INITIAL_INTEREST_RATE_CAP", "PERIODIC_INTEREST_RATE_CAP",
                        "LIFETIME_INTEREST_RATE_CAP", "MARGIN", "BALLOON_INDICATOR",
                        "PLAN_NUMBER", "FORBEARANCE_INDICATOR", "HIGH_LOAN_TO_VALUE_HLTV_REFINANCE_OPTION_INDICATOR",
                        "DEAL_NAME", "RE_PROCS_FLAG", "ADR_TYPE", "ADR_COUNT", "ADR_UPB", "PAYMENT_DEFERRAL_MOD_EVENT_FLAG", "INTEREST_BEARING_UPB")

lppub_column_names = tolower(lppub_column_names)

lppub_column_classes <- c("character", "character", "character", "character", "character", "character",
                          "character", "numeric", "numeric", "numeric", "numeric",
                          "numeric", "numeric", "character", "character", "numeric", "numeric",
                          "numeric", "character", "numeric", "numeric", "character", "numeric",
                          "numeric", "numeric", "character", "character", "character",
                          "numeric", "character", "character", "character", "character",
                          "numeric", "character", "character", "character", "character",
                          "numeric", "character", "character", "character", "character",
                          "character", "character", "numeric", "character", "numeric",
                          "numeric", "numeric", "character", "character", "character",
                          "numeric", "numeric", "numeric", "numeric", "numeric", "numeric",
                          "numeric", "numeric", "numeric", "numeric", "numeric", "character",
                          "numeric", "character", "numeric", "numeric", "numeric", "numeric",
                          "numeric", "numeric", "character", "numeric", "numeric", "numeric",
                          "numeric", "character", "numeric", "character", "numeric", "character",
                          "numeric", "numeric", "character", "character", "numeric", "numeric",
                          "numeric", "numeric", "numeric", "numeric", "numeric", "numeric",
                          "numeric", "numeric", "numeric", "numeric", "numeric", "character",
                          "character", "character", "character", "character",
                          "character", "numeric", "numeric", "character", "numeric")

source("scrap.R")

# the data contained all the data from 
system.time(fm.df <- zip_to_disk.frame(
  "c:/data/fannie-mae-single-family/Performance_All.zip",
  outdir = "c:/data/fannie-mae-disk.frame/", 
  # passed to fread
  col.names = lppub_column_names,
  colClasses = lppub_column_classes,
  shardby = "loan_id"
  ))

system.time(a <- list.dirs("c:/data/fannie-mae-disk.frame/", recursive = F) %>% 
  lapply(disk.frame) %>% 
  rbindlist.disk.frame(outdir = "c:/data/fannie-mae-disk.frame2/"))

a = disk.frame("c:/data/fannie-mae-disk.frame2/")
head(a)

system.time(b <- a %>% 
              group_by(pool_id) %>% 
              summarise(n()) %>% 
              collect)
duckdb / duckdb-r

R: Add support for the `type`/`dtypes` in `duckdb_read_csv`? #383
