diff --git a/base/db/.Rbuildignore b/base/db/.Rbuildignore
index 91114bf2f2b..02548d35d12 100644
--- a/base/db/.Rbuildignore
+++ b/base/db/.Rbuildignore
@@ -1,2 +1,3 @@
 ^.*\.Rproj$
 ^\.Rproj\.user$
+try\.sqlite
diff --git a/base/db/.gitignore b/base/db/.gitignore
new file mode 100644
index 00000000000..c1e2f2a3c7b
--- /dev/null
+++ b/base/db/.gitignore
@@ -0,0 +1,2 @@
+try.sqlite
+inst/import-try/data-proc
diff --git a/base/db/DESCRIPTION b/base/db/DESCRIPTION
index 0e1d9183067..82366762c69 100644
--- a/base/db/DESCRIPTION
+++ b/base/db/DESCRIPTION
@@ -18,6 +18,10 @@ Imports:
     PEcAn.utils,
     dbplyr (>= 1.2.0),
     dplyr,
+    tibble,
+    purrr,
+    tidyr,
+    glue,
     lubridate,
     magrittr,
     ncdf4,
@@ -25,7 +29,12 @@ Imports:
     udunits2
 Suggests:
     RPostgreSQL,
-    testthat (>= 1.0.2)
+    RSQLite,
+    testthat (>= 1.0.2),
+    tidyverse,
+    data.table,
+    rcrossref,
+    here
 License: FreeBSD + file LICENSE
 Copyright: Authors
 LazyLoad: yes
diff --git a/base/db/NAMESPACE b/base/db/NAMESPACE
index 78717a7f0b4..01bcbcd748e 100644
--- a/base/db/NAMESPACE
+++ b/base/db/NAMESPACE
@@ -1,5 +1,6 @@
 # Generated by roxygen2: do not edit by hand
 
+export("%>%")
 export(append.covariate)
 export(assign.treatments)
 export(bety2pecan)
@@ -12,6 +13,7 @@ export(db.print.connections)
 export(db.query)
 export(db.showQueries)
 export(dbHostInfo)
+export(db_merge_into)
 export(dbfile.check)
 export(dbfile.file)
 export(dbfile.id)
@@ -32,7 +34,9 @@ export(get_run_ids)
 export(get_users)
 export(get_var_names)
 export(get_workflow_ids)
+export(insert_table)
 export(load_data_single_run)
+export(match_dbcols)
 export(ncdays2date)
 export(query.file.path)
 export(query.format.vars)
@@ -43,7 +47,9 @@ export(query.trait.data)
 export(query.traits)
 export(rename_jags_columns)
 export(runs)
+export(search_references)
 export(take.samples)
+export(try2sqlite)
 export(var_names_all)
 export(workflow)
 export(workflows)
diff --git a/base/db/R/db_merge_into.R b/base/db/R/db_merge_into.R
new file mode 100644
index 00000000000..a00718e5ac3
--- /dev/null
+++ b/base/db/R/db_merge_into.R
@@ -0,0 +1,29 @@
+#' Merge local data frame into SQL table
+#'
+#' @inheritParams insert_table
+#' @inheritDotParams insert_table
+#' @param by Character vector of columns by which to perform merge. Defaults to all columns in `values`
+#' @return Data frame: Inner join of SQL table and input data frame (as unevaluated "lazy query" table)
+#' @export
+#' @examples
+#' irisdb <- DBI::dbConnect(RSQLite::SQLite(), ":memory:")
+#' dplyr::copy_to(irisdb, iris[1:10,], name = "iris", overwrite = TRUE)
+#' db_merge_into(iris[1:12,], "iris", irisdb)
+#' dplyr::tbl(irisdb, "iris") %>% dplyr::count()
+db_merge_into <- function(values, table, con, by = NULL, drop = FALSE, ...) {
+  values_fixed <- match_dbcols(values, table, con, drop = FALSE)
+  if (is.null(by)) {
+    by <- match_colnames(values, table, con)
+  }
+  sql_tbl <- dplyr::tbl(con, table)
+  values_merge <- dplyr::anti_join(values_fixed, sql_tbl, by = by, copy = TRUE)
+  if (nrow(values_merge) < 1 || ncol(values_merge) < 1) {
+    PEcAn.logger::logger.warn(
+      "Input table for merge is empty."
+    )
+  } else {
+    insert <- insert_table(values_merge, table, con, ...)
+  }
+  dplyr::tbl(con, table) %>%
+    dplyr::inner_join(values_fixed, copy = TRUE)
+}
diff --git a/base/db/R/insert_table.R b/base/db/R/insert_table.R
new file mode 100644
index 00000000000..f9e45c488bf
--- /dev/null
+++ b/base/db/R/insert_table.R
@@ -0,0 +1,112 @@
+#' Insert R data frame into SQL database
+#'
+#' First, subset to matching columns. Then, make sure the local and SQL column 
+#' classes match, coercing local to SQL as necessary (or throwing an error). 
+#' Then, build an SQL string for the insert statement. Finally, insert into the 
+#' database.
+#'
+#' @param values `data.frame` of values to write to SQL database
+#' @param table Name of target SQL table, as character
+#' @param coerce_col_class logical, whether or not to coerce local data columns 
+#' to SQL classes. Default = `TRUE.`
+#' @param drop logical. If `TRUE` (default), drop columns not found in SQL table.
+#' @inheritParams db.query
+#' @inherit db.query return
+#' @export
+#' @examples
+#' irisdb <- DBI::dbConnect(RSQLite::SQLite(), ":memory:")
+#' dplyr::copy_to(irisdb, iris[1,], name = "iris", overwrite = TRUE)
+#' insert_table(iris[-1,], "iris", irisdb$con)
+#' dplyr::tbl(irisdb, "iris")
+insert_table <- function(values, table, con, coerce_col_class = TRUE, drop = TRUE) {
+  values_fixed <- match_dbcols(values, table, con, coerce_col_class, drop = TRUE)
+  insert_query <- build_insert_query(values_fixed, table, .con = con)
+  db.query(insert_query, con)
+}
+
+#' Match column names and classes between local and SQL table
+#'
+#' @inheritParams insert_table
+#' @return `values` `data.frame` with column names and classes matched to SQL
+#' @export
+match_dbcols <- function(values, table, con, coerce_col_class = TRUE, drop = TRUE) {
+  use_cols <- match_colnames(values, table, con)
+  if (length(use_cols) < 1) {
+    PEcAn.logger::logger.severe(
+      "No columns match between input and target table."
+    )
+  }
+  PEcAn.logger::logger.debug(
+    "Matched the following cols: ",
+    paste(use_cols, collapse = ", ")
+  )
+  values_sub <- values[, use_cols]
+  # Load one row to get column types
+  sql_row <- dplyr::tbl(con, table) %>% head(1) %>% dplyr::collect()
+  sql_types <- purrr::map(sql_row, class) %>%
+    purrr::map_chr(1) %>%
+    .[use_cols]
+  values_types <- purrr::map(values_sub, class) %>% purrr::map_chr(1)
+  type_mismatch <- sql_types != values_types
+  if (sum(type_mismatch) > 0) {
+    mismatch_string <- sprintf(
+      "%s: local is %s, SQL is %s",
+      names(values_types),
+      values_types,
+      sql_types
+    )[type_mismatch]
+    PEcAn.logger::logger.info(
+      "Found type mismatches in the following columns: ",
+      paste0(mismatch_string, collapse = "; ")
+    )
+    if (!coerce_col_class) {
+      PEcAn.logger::logger.severe(
+        "Type mismatch detected, and `coerce_col_class` is `FALSE`. ",
+        "Fix column class mismatches manually."
+      )
+    } else {
+      PEcAn.logger::logger.info(
+        "Coercing local column types to match SQL."
+      )
+      # Coerce values data frame to these types
+      values_fixed <- purrr::map2_dfc(values_sub, sql_types, as)
+    }
+  } else {
+    values_fixed <- values_sub
+  }
+  if (drop) {
+    values_fixed
+  } else {
+    drop_cols <- colnames(values)[!colnames(values) %in% use_cols]
+    dplyr::bind_cols(values_fixed, values[, drop_cols])
+  }
+}
+
+#' Match names of local data frame to SQL table
+#'
+#' @inheritParams insert_table
+match_colnames <- function(values, table, con) {
+  tbl_db <- dplyr::tbl(con, table)
+  table_cols <- dplyr::tbl_vars(tbl_db)
+  values_cols <- colnames(values)
+  intersect(values_cols, table_cols)
+}
+
+#' Build query to insert R data frame into SQL table
+#'
+#' @inheritParams insert_table
+#' @inheritParams glue::glue_sql
+build_insert_query <- function(values, table, .con) {
+  value_list <- purrr::map(seq_len(nrow(values)), ~as.list(values[.x, ]))
+
+  insert_list <- value_list %>%
+    purrr::map(unname) %>%
+    purrr::map(dbplyr::escape) %>%
+    purrr::map(dbplyr::sql_vector)
+
+  glue::glue_sql(
+    "INSERT INTO {`table`} ({`colnames(values)`*}) ",
+    "VALUES {insert_list*}",
+    .con = .con
+  )
+}
diff --git a/base/db/R/search_references.R b/base/db/R/search_references.R
new file mode 100644
index 00000000000..14b61adc7bb
--- /dev/null
+++ b/base/db/R/search_references.R
@@ -0,0 +1,67 @@
+#' Perform crossref search for a list of references
+#'
+#' @param queries Character vector of queries
+#' @inheritDotParams search_reference_single
+#' @inherit search_reference_single description return
+#' @export
+search_references <- function(queries, ...) {
+  search_fun <- search_reference_single %>%
+    purrr::partial(...) %>%
+    purrr::possibly(otherwise = data.frame(title = "Not found"))
+  encodeString(queries) %>%
+    purrr::map_dfr(search_fun)
+}
+
+#' Perform crossref search for a single reference
+#'
+#' Requires the `rcrossref` package.
+#'
+#' @param query Citation string (length 1) to search for DOI
+#' @param min_score Minimum match score. Default (85) is fairly strict.
+#' @param limit Number of results to return
+#' @return `data.frame` containing crossref information converted to match bety citations table.
+search_reference_single <- function(query, limit = 1, min_score = 85) {
+  stopifnot(length(query) == 1)
+  PEcAn.logger::logger.debug("Processing query:\n", query)
+  crsearch <- rcrossref::cr_works(query = query, limit = limit)
+  if (is.null(crsearch[["data"]])) {
+    PEcAn.logger::logger.warn(
+      "Error in crossref query. ",
+      "Setting title to search string and leaving other fields blank."
+    )
+    return(tibble::tibble(query = query))
+  }
+  crdata <- crsearch[["data"]] %>%
+    dplyr::mutate(score = as.numeric(score)) %>%
+    dplyr::filter(score > min_score)
+  if (nrow(crdata) < 1) {
+    PEcAn.logger::logger.info(
+      "No matches found. ",
+      "Setting title to search string and leaving other fields blank.")
+    return(tibble::tibble(query = query))
+  }
+  keep_cols <- c(
+    "author",
+    "year",
+    "title",
+    journal = "container.title",
+    vol = "volume",
+    pg = "page",
+    doi = "DOI",
+    "score",
+    "query"
+  )
+  proc_search <- crdata %>%
+    dplyr::mutate(
+      # Get the first author only -- this is the BETY format
+      author_family = purrr::map(author, list("family", 1)),
+      author_given = purrr::map(author, list("given", 1)),
+      author = paste(author_family, author_given, sep = ", "),
+      year = gsub("([[:digit:]]{4}).*", "\\1", issued) %>% as.numeric(),
+      query = query,
+      score = as.numeric(score)
+    )
+  use_cols <- keep_cols[keep_cols %in% colnames(proc_search)]
+  dplyr::select(proc_search, !!!use_cols)
+}
+
diff --git a/base/db/R/try2sqlite.R b/base/db/R/try2sqlite.R
new file mode 100644
index 00000000000..828afd00957
--- /dev/null
+++ b/base/db/R/try2sqlite.R
@@ -0,0 +1,91 @@
+#' Convert TRY text file to SQLite database
+#'
+#' The TRY file is huge and unnecessarily long, which makes it difficult to 
+#' work with. The resulting SQLite database is much smaller on disk, and can be 
+#' read much faster thanks to lazy evaluation.
+#'
+#' The resulting TRY SQLite database contains the following tables:
+#'  - `values` -- The actual TRY data. Links to all other tables through ID columns.
+#'  - `traits` -- Description of trait and data names. Links to `values` through `DataID`. Similar to BETY `variables` table.
+#'  - `datasets` -- Description of datasets and references/citations. Links to `values` through `DatasetID` and `ReferenceID`.
+#'  - `species` -- Species. Links to `values` through `AccSpeciesID`.
+#'
+#' @param try_files Character vector of file names containing TRY data.  
+#' Multiple files are combined with `data.table::rbindlist`.
+#' @param sqlite_file Target SQLite database file name, as character.
+#' @export
+try2sqlite <- function(try_files, sqlite_file = "try.sqlite") {
+  # Read files
+  PEcAn.logger::logger.info("Reading in TRY data...")
+  raw_data <- Map(data.table::fread, try_files) %>%
+    data.table::rbindlist()
+
+  # Create integer reference ID for compact storage
+  PEcAn.logger::logger.info("Adding ReferenceID column")
+  raw_data[["ReferenceID"]] <- as.integer(factor(raw_data[["Reference"]]))
+
+  # Create tables
+  PEcAn.logger::logger.info("Extracting data values table.")
+  data_cols <- c(
+    "ObsDataID",        # TRY row ID -- unique to each observation of a given trait
+    "ObservationID",    # TRY "entity" ID -- identifies a set of trait measurements (e.g. leaf)
+    "DataID",           # Links to data ID
+    "StdValue",         # Standardized, QA-QC'ed value
+    "UnitName",         # Standardized unit
+    "AccSpeciesID",     # Link to 'species' table
+    "DatasetID",        # Link to 'datasets' table.
+    "ReferenceID",      # Link to 'try_references' table.
+    "ValueKindName",    # Type of value, e.g. mean, min, max, etc.
+    "UncertaintyName",  # Kind of uncertainty
+    "Replicates",       # Number of replicates
+    "RelUncertaintyPercent",
+    "OrigValueStr",         # Original data, as character string (before QA/QC)
+    "OrigUnitStr",          # Original unit, as character string (before QA/QC)
+    "OrigUncertaintyStr"    # Original uncertainty, as character string (before QA/QC)
+  )
+  data_values <- unique(raw_data[, data_cols, with = FALSE])
+
+  PEcAn.logger::logger.info("Extrating datasets table...")
+  datasets_cols <- c(
+    "DatasetID",
+    "Dataset",
+    "LastName",
+    "FirstName",
+    "Reference",
+    "ReferenceID"
+  )
+  datasets_values <- unique(raw_data[, datasets_cols, with = FALSE])
+
+  PEcAn.logger::logger.info("Extracting traits table...")
+  traits_cols <- c(
+    "DataID",
+    "DataName",
+    "TraitID",
+    "TraitName"
+  )
+  traits_values <- unique(raw_data[, traits_cols, with = FALSE])
+
+  PEcAn.logger::logger.info("Extracting species table...")
+  species_cols <- c(
+    "AccSpeciesID",
+    "AccSpeciesName",
+    "SpeciesName"
+  )
+  species_values <- unique(raw_data[, species_cols, with = FALSE])
+
+  PEcAn.logger::logger.info("Writing tables to SQLite database...")
+  con <- DBI::dbConnect(RSQLite::SQLite(), sqlite_file)
+  on.exit(DBI::dbDisconnect(con))
+  PEcAn.logger::logger.info("Writing values table...")
+  DBI::dbWriteTable(con, "values", data_values)
+  PEcAn.logger::logger.info("Writing traits table...")
+  DBI::dbWriteTable(con, "traits", traits_values)
+  PEcAn.logger::logger.info("Writing datasets table...")
+  DBI::dbWriteTable(con, "datasets", datasets_values)
+  PEcAn.logger::logger.info("Writing species table...")
+  DBI::dbWriteTable(con, "species", species_values)
+
+  PEcAn.logger::logger.info("Done creating TRY SQLite database!")
+
+  NULL
+}
diff --git a/base/db/R/zz.imports.R b/base/db/R/zz.imports.R
index 050b5a37cf9..4eb0423a534 100644
--- a/base/db/R/zz.imports.R
+++ b/base/db/R/zz.imports.R
@@ -1,3 +1,9 @@
-##' Imports from other packages
-##'
-##' @importFrom magrittr `%>%`
+#' Imports from other packages
+#'
+#' @name otherimports
+#' @importFrom magrittr %>%
+#' @export
+magrittr::`%>%`
+
+#' @rdname otherimports
+#' @importFrom rlang !! !!!
diff --git a/base/db/inst/import-try/01_try_sqlite.R b/base/db/inst/import-try/01_try_sqlite.R
new file mode 100644
index 00000000000..9480ea8687e
--- /dev/null
+++ b/base/db/inst/import-try/01_try_sqlite.R
@@ -0,0 +1,18 @@
+# Create the TRY SQLite database
+library(PEcAn.DB)
+library(PEcAn.logger)
+library(here)
+
+configfile <- here("inst", "import-try", "config.R")
+source(configfile)
+
+if (!exists("overwrite")) {
+  overwrite <- FALSE
+}
+
+if (!file.exists(sqlite_file) || overwrite) {
+  file.remove(sqlite_file)
+  try2sqlite(try_files, sqlite_file)
+} else {
+  logger.info("TRY SQLite database already exists and `overwrite` is FALSE. ")
+}
diff --git a/base/db/inst/import-try/02_citations.R b/base/db/inst/import-try/02_citations.R
new file mode 100644
index 00000000000..98c1dc56758
--- /dev/null
+++ b/base/db/inst/import-try/02_citations.R
@@ -0,0 +1,71 @@
+# Add DOIs to TRY citations
+library(tidyverse)
+library(DBI)
+library(RSQLite)
+library(rcrossref)
+library(PEcAn.logger)
+library(PEcAn.DB)
+library(here)
+
+wd <- here("inst", "import-try")
+configfile <- file.path(wd, "config.R")
+source(configfile)
+
+data_dir <- file.path(wd, "data-proc")
+dir.create(data_dir, showWarnings = FALSE)
+
+bety <- db.open(betyparams)
+if (!"notes" %in% dbListFields(bety, "citations")) {
+  logger.severe(
+    "`notes` column required in Bety citations table ",
+    "but not found in this version of Bety. ",
+    "Please make sure you have performed schema migration ",
+    "version 20180206152600 (relax_citations)."
+  )
+}
+
+trydb <- dbConnect(SQLite(), sqlite_file)
+
+reference_dat <- tbl(trydb, "datasets") %>%
+  distinct(Reference, ReferenceID) %>%
+  collect()
+
+refs_proc_file <- file.path(data_dir, "refs_proc.rds")
+if (file.exists(refs_proc_file)) {
+  refs_proc <- readRDS(refs_proc_file)
+} else {
+  logger.setLevel("DEBUG")    # To get status messages
+  refs_proc <- reference_dat %>%
+    mutate(cr_df = map(Reference, search_references, min_score = 40)) %>%
+    unnest()
+  logger.setLevel("INFO")
+  saveRDS(refs_proc, refs_proc_file)
+}
+
+# Replace bad matches with NA
+minscore <- 85
+fill_na <- function(field, score) {
+  na <- as(NA, class(field))
+  if_else(score > minscore, field, na)
+}
+refs_proc2 <- refs_proc %>%
+  mutate_at(
+    c("title", "author", "year", "journal", "vol", "pg", "doi"),
+    fill_na,
+    score = .$score
+  ) %>%
+  mutate(
+    title = if_else(!is.na(title), title, paste0("TRY ReferenceID ", ReferenceID)),
+    author = if_else(!is.na(author), author, "Unknown TRY data (see title)"),
+    author = substr(author, 0, 254),    # Trim author to 255 characters
+    journal = if_else(!is.na(journal), journal, "Unknown TRY data (see title)"),
+    # Use the Kattge 2007 TRY paper's DOI as a placeholder
+    doi = if_else(!is.na(doi), doi, "10.1111/j.1365-2486.2011.02451.x"),
+    year = if_else(!is.na(year), year, 2018),
+    pg = if_else(!is.na(pg), pg, "9999"),
+    notes = paste("Original TRY reference: ", Reference)
+  )
+
+bety_refs <- db_merge_into(refs_proc2, "citations", bety, "notes") %>%
+  collect()
+saveRDS(bety_refs, file.path(data_dir, "refs_bety.rds"))
diff --git a/base/db/inst/import-try/01.global.subset.R b/base/db/inst/import-try/91.global.subset.R
similarity index 100%
rename from base/db/inst/import-try/01.global.subset.R
rename to base/db/inst/import-try/91.global.subset.R
diff --git a/base/db/inst/import-try/02.data.specific.subset.R b/base/db/inst/import-try/92.data.specific.subset.R
similarity index 100%
rename from base/db/inst/import-try/02.data.specific.subset.R
rename to base/db/inst/import-try/92.data.specific.subset.R
diff --git a/base/db/inst/import-try/03.create.try.sites.R b/base/db/inst/import-try/93.create.try.sites.R
similarity index 100%
rename from base/db/inst/import-try/03.create.try.sites.R
rename to base/db/inst/import-try/93.create.try.sites.R
diff --git a/base/db/inst/import-try/04.match.species.R b/base/db/inst/import-try/94.match.species.R
similarity index 100%
rename from base/db/inst/import-try/04.match.species.R
rename to base/db/inst/import-try/94.match.species.R
diff --git a/base/db/inst/import-try/05.citations.R b/base/db/inst/import-try/95.citations.R
similarity index 100%
rename from base/db/inst/import-try/05.citations.R
rename to base/db/inst/import-try/95.citations.R
diff --git a/base/db/inst/import-try/06.load.data.R b/base/db/inst/import-try/96.load.data.R
similarity index 100%
rename from base/db/inst/import-try/06.load.data.R
rename to base/db/inst/import-try/96.load.data.R
diff --git a/base/db/inst/import-try/README.md b/base/db/inst/import-try/README.md
index 33ec4ade800..7ad5b00de76 100644
--- a/base/db/inst/import-try/README.md
+++ b/base/db/inst/import-try/README.md
@@ -1,8 +1,13 @@
 ---
-title: "Import TRY database into BETY""
+title: "Import TRY database into BETY"
 Author: "Alexey Shiklomanov"
 ---
 
+# Workflow
+
+1. `01_try_sqlite.R` -- Read TRY data files (huge tab-delimited text files) and convert them to an SQLite database, which is much faster and easier to work with.
+2. `02_citations.R` -- Add DOIs to TRY references, if possible.
+
 # Package dependencies
 1. `data.table` -- Makes it remotely possible to work with the TRY database. Requires an up-to-date version, so if parts of the workflow break, try re-installing this.
 2. `bit64` -- Used by `data.table` to read and store large integers, which constitude most of the ID's in TRY and BETY.
diff --git a/base/db/inst/import-try/bety_connect.R b/base/db/inst/import-try/bety_connect.R
new file mode 100644
index 00000000000..89d20b37f12
--- /dev/null
+++ b/base/db/inst/import-try/bety_connect.R
@@ -0,0 +1,6 @@
+bety <- PEcAn.DB::db.open(list(
+  host = "localhost",
+  dbname = "bety",
+  user = "bety",
+  password = "bety"
+))
diff --git a/base/db/inst/import-try/config.R b/base/db/inst/import-try/config.R
new file mode 100644
index 00000000000..eff2c8ce7a1
--- /dev/null
+++ b/base/db/inst/import-try/config.R
@@ -0,0 +1,14 @@
+# Character vector containing TRY file paths
+try_files <- c("~/Projects/try-data/4143.txt")
+stopifnot(all(file.exists(try_files)))
+
+# Path to generated TRY SQLite file
+sqlite_file <- "inst/import-try/try.sqlite"
+
+# Bety connection configuration
+betyparams <- list(
+  host = "localhost",
+  dbname = "bety",
+  user = "bety",
+  password = "bety"
+)
diff --git a/base/db/man/build_insert_query.Rd b/base/db/man/build_insert_query.Rd
new file mode 100644
index 00000000000..bbfb23fcfd4
--- /dev/null
+++ b/base/db/man/build_insert_query.Rd
@@ -0,0 +1,18 @@
+% Generated by roxygen2: do not edit by hand
+% Please edit documentation in R/insert_table.R
+\name{build_insert_query}
+\alias{build_insert_query}
+\title{Build query to insert R data frame into SQL table}
+\usage{
+build_insert_query(values, table, .con)
+}
+\arguments{
+\item{values}{`data.frame` of values to write to SQL database}
+
+\item{table}{Name of target SQL table, as character}
+
+\item{.con}{[\code{DBIConnection}]:A DBI connection object obtained from \code{DBI::dbConnect()}.}
+}
+\description{
+Build query to insert R data frame into SQL table
+}
diff --git a/base/db/man/db_merge_into.Rd b/base/db/man/db_merge_into.Rd
new file mode 100644
index 00000000000..63752d4b027
--- /dev/null
+++ b/base/db/man/db_merge_into.Rd
@@ -0,0 +1,41 @@
+% Generated by roxygen2: do not edit by hand
+% Please edit documentation in R/db_merge_into.R
+\name{db_merge_into}
+\alias{db_merge_into}
+\title{Merge local data frame into SQL table}
+\usage{
+db_merge_into(values, table, con, by = NULL, drop = FALSE, ...)
+}
+\arguments{
+\item{values}{`data.frame` of values to write to SQL database}
+
+\item{table}{Name of target SQL table, as character}
+
+\item{con}{database connection object}
+
+\item{by}{Character vector of columns by which to perform merge. Defaults to all columns in `values`}
+
+\item{drop}{logical. If `TRUE` (default), drop columns not found in SQL table.}
+
+\item{...}{Arguments passed on to \code{insert_table}
+\describe{
+  \item{values}{`data.frame` of values to write to SQL database}
+  \item{table}{Name of target SQL table, as character}
+  \item{coerce_col_class}{logical, whether or not to coerce local data columns 
+to SQL classes. Default = `TRUE.`}
+  \item{drop}{logical. If `TRUE` (default), drop columns not found in SQL table.}
+  \item{con}{database connection object}
+}}
+}
+\value{
+Data frame: Inner join of SQL table and input data frame (as unevaluated "lazy query" table)
+}
+\description{
+Merge local data frame into SQL table
+}
+\examples{
+irisdb <- DBI::dbConnect(RSQLite::SQLite(), ":memory:")
+dplyr::copy_to(irisdb, iris[1:10,], name = "iris", overwrite = TRUE)
+db_merge_into(iris[1:12,], "iris", irisdb)
+dplyr::tbl(irisdb, "iris") \%>\% dplyr::count()
+}
diff --git a/base/db/man/insert_table.Rd b/base/db/man/insert_table.Rd
new file mode 100644
index 00000000000..c51740868ce
--- /dev/null
+++ b/base/db/man/insert_table.Rd
@@ -0,0 +1,35 @@
+% Generated by roxygen2: do not edit by hand
+% Please edit documentation in R/insert_table.R
+\name{insert_table}
+\alias{insert_table}
+\title{Insert R data frame into SQL database}
+\usage{
+insert_table(values, table, con, coerce_col_class = TRUE, drop = TRUE)
+}
+\arguments{
+\item{values}{`data.frame` of values to write to SQL database}
+
+\item{table}{Name of target SQL table, as character}
+
+\item{con}{database connection object}
+
+\item{coerce_col_class}{logical, whether or not to coerce local data columns 
+to SQL classes. Default = `TRUE.`}
+
+\item{drop}{logical. If `TRUE` (default), drop columns not found in SQL table.}
+}
+\value{
+data frame with query results
+}
+\description{
+First, subset to matching columns. Then, make sure the local and SQL column 
+classes match, coercing local to SQL as necessary (or throwing an error). 
+Then, build an SQL string for the insert statement. Finally, insert into the 
+database.
+}
+\examples{
+irisdb <- DBI::dbConnect(RSQLite::SQLite(), ":memory:")
+dplyr::copy_to(irisdb, iris[1,], name = "iris", overwrite = TRUE)
+insert_table(iris[-1,], "iris", irisdb$con)
+dplyr::tbl(irisdb, "iris")
+}
diff --git a/base/db/man/match_colnames.Rd b/base/db/man/match_colnames.Rd
new file mode 100644
index 00000000000..38450a89bfa
--- /dev/null
+++ b/base/db/man/match_colnames.Rd
@@ -0,0 +1,18 @@
+% Generated by roxygen2: do not edit by hand
+% Please edit documentation in R/insert_table.R
+\name{match_colnames}
+\alias{match_colnames}
+\title{Match names of local data frame to SQL table}
+\usage{
+match_colnames(values, table, con)
+}
+\arguments{
+\item{values}{`data.frame` of values to write to SQL database}
+
+\item{table}{Name of target SQL table, as character}
+
+\item{con}{database connection object}
+}
+\description{
+Match names of local data frame to SQL table
+}
diff --git a/base/db/man/match_dbcols.Rd b/base/db/man/match_dbcols.Rd
new file mode 100644
index 00000000000..ba0b6ae8d00
--- /dev/null
+++ b/base/db/man/match_dbcols.Rd
@@ -0,0 +1,26 @@
+% Generated by roxygen2: do not edit by hand
+% Please edit documentation in R/insert_table.R
+\name{match_dbcols}
+\alias{match_dbcols}
+\title{Match column names and classes between local and SQL table}
+\usage{
+match_dbcols(values, table, con, coerce_col_class = TRUE, drop = TRUE)
+}
+\arguments{
+\item{values}{`data.frame` of values to write to SQL database}
+
+\item{table}{Name of target SQL table, as character}
+
+\item{con}{database connection object}
+
+\item{coerce_col_class}{logical, whether or not to coerce local data columns 
+to SQL classes. Default = `TRUE.`}
+
+\item{drop}{logical. If `TRUE` (default), drop columns not found in SQL table.}
+}
+\value{
+`values` `data.frame` with column names and classes matched to SQL
+}
+\description{
+Match column names and classes between local and SQL table
+}
diff --git a/base/db/man/otherimports.Rd b/base/db/man/otherimports.Rd
new file mode 100644
index 00000000000..dd79f7fbbc3
--- /dev/null
+++ b/base/db/man/otherimports.Rd
@@ -0,0 +1,19 @@
+% Generated by roxygen2: do not edit by hand
+% Please edit documentation in R/zz.imports.R
+\docType{import}
+\name{otherimports}
+\alias{otherimports}
+\alias{\%>\%}
+\title{Imports from other packages}
+\description{
+Imports from other packages
+}
+\keyword{internal}
+\description{
+These objects are imported from other packages. Follow the links
+below to see their documentation.
+
+\describe{
+  \item{magrittr}{\code{\link[magrittr]{\%>\%}}}
+}}
+
diff --git a/base/db/man/search_reference_single.Rd b/base/db/man/search_reference_single.Rd
new file mode 100644
index 00000000000..c689ccdb86d
--- /dev/null
+++ b/base/db/man/search_reference_single.Rd
@@ -0,0 +1,21 @@
+% Generated by roxygen2: do not edit by hand
+% Please edit documentation in R/search_references.R
+\name{search_reference_single}
+\alias{search_reference_single}
+\title{Perform crossref search for a single reference}
+\usage{
+search_reference_single(query, limit = 1, min_score = 85)
+}
+\arguments{
+\item{query}{Citation string (length 1) to search for DOI}
+
+\item{limit}{Number of results to return}
+
+\item{min_score}{Minimum match score. Default (85) is fairly strict.}
+}
+\value{
+`data.frame` containing crossref information converted to match bety citations table.
+}
+\description{
+Requires the `rcrossref` package.
+}
diff --git a/base/db/man/search_references.Rd b/base/db/man/search_references.Rd
new file mode 100644
index 00000000000..8af809bf7a1
--- /dev/null
+++ b/base/db/man/search_references.Rd
@@ -0,0 +1,24 @@
+% Generated by roxygen2: do not edit by hand
+% Please edit documentation in R/search_references.R
+\name{search_references}
+\alias{search_references}
+\title{Perform crossref search for a list of references}
+\usage{
+search_references(queries, ...)
+}
+\arguments{
+\item{queries}{Character vector of queries}
+
+\item{...}{Arguments passed on to \code{search_reference_single}
+\describe{
+  \item{query}{Citation string (length 1) to search for DOI}
+  \item{min_score}{Minimum match score. Default (85) is fairly strict.}
+  \item{limit}{Number of results to return}
+}}
+}
+\value{
+`data.frame` containing crossref information converted to match bety citations table.
+}
+\description{
+Perform crossref search for a list of references
+}
diff --git a/base/db/man/try2sqlite.Rd b/base/db/man/try2sqlite.Rd
new file mode 100644
index 00000000000..e4b6407712b
--- /dev/null
+++ b/base/db/man/try2sqlite.Rd
@@ -0,0 +1,26 @@
+% Generated by roxygen2: do not edit by hand
+% Please edit documentation in R/try2sqlite.R
+\name{try2sqlite}
+\alias{try2sqlite}
+\title{Convert TRY text file to SQLite database}
+\usage{
+try2sqlite(try_files, sqlite_file = "try.sqlite")
+}
+\arguments{
+\item{try_files}{Character vector of file names containing TRY data.
+Multiple files are combined with `data.table::rbindlist`.}
+
+\item{sqlite_file}{Target SQLite database file name, as character.}
+}
+\description{
+The TRY file is huge and unnecessarily long, which makes it difficult to 
+work with. The resulting SQLite database is much smaller on disk, and can be 
+read much faster thanks to lazy evaluation.
+}
+\details{
+The resulting TRY SQLite database contains the following tables:
+ - `values` -- The actual TRY data. Links to all other tables through ID columns.
+ - `traits` -- Description of trait and data names. Links to `values` through `DataID`. Similar to BETY `variables` table.
+ - `datasets` -- Description of datasets and references/citations. Links to `values` through `DatasetID` and `ReferenceID`.
+ - `species` -- Species. Links to `values` through `AccSpeciesID`.
+}
diff --git a/base/db/tests/testthat/test.insert.R b/base/db/tests/testthat/test.insert.R
new file mode 100644
index 00000000000..95dd80a459e
--- /dev/null
+++ b/base/db/tests/testthat/test.insert.R
@@ -0,0 +1,49 @@
+library(PEcAn.DB)
+library(testthat)
+context("SQL insertion helper functions")
+
+test_that(
+  "RSQLite-dependent tests work",
+  {
+    skip_if_not_installed("RSQLite")
+    irisdb <- DBI::dbConnect(RSQLite::SQLite(), ":memory:")
+    iris <- tibble::as_tibble(iris) %>%
+      dplyr::mutate(
+        Species = as.character(Species)
+      )
+    dplyr::copy_to(irisdb, iris[1,], "iris", overwrite = TRUE)
+    # Add extra column to see if it's successfully ignored
+    iris2 <- dplyr::mutate(iris, extracol = seq_len(nrow(iris)))
+    iris_insert <- iris2[2:10,]
+    .insert <- insert_table(iris_insert, "iris", irisdb)
+    test_that(
+      "Subset of iris was inserted into database",
+      {
+        iris_insert_test <- dplyr::tbl(irisdb, "iris") %>% dplyr::collect()
+        expect_equal(iris[1:10,], iris_insert_test)
+      }
+    )
+
+    iris_merge <- iris2[5:12,]
+    out_merge <- db_merge_into(iris_merge, "iris", irisdb) %>%
+      dplyr::collect()
+    iris_merge_nrow <- dplyr::tbl(irisdb, "iris") %>%
+      dplyr::count() %>%
+      dplyr::pull(n)
+    test_that(
+      "Only subset of iris data were merged",
+      {
+        expect_equal(out_merge, iris2[5:12,])
+        out_merge2 <- db_merge_into(iris_merge, "iris", irisdb) %>%
+          dplyr::collect()
+        expect_equal(out_merge, out_merge2)
+      }
+    )
+
+    test_that(
+      "Extra column (not in SQL) was retained in `out_merge`",
+      {
+        expect_true("extracol" %in% colnames(out_merge))
+      }
+    )
+  })