Skip to content

Configuration

All options are passed to require('vim-teradata').setup({ ... }). All fields are optional except users.

require('vim-teradata').setup({
  users = {
    { log_mech = 'TD2', user = 'USER', tdpid = 'HOST' },
  },
  -- optional settings shown with their defaults:
  retlimit = 100,
  replacements = {},
  ft = { 'sql', 'teradata' },
  picker = 'auto',
  sql_detection = 'always',
  timeout_ms = nil,
  history_max = nil,
  sep = '~',
  edit_enabled = true,
  preview_updates = true,
  null_token = 'NULL',
  edit_keymaps = { toggle = 'E', save = 'S', cancel = 'C' },
  rerun_keymaps = { sql = 'g.', result = '.' },
  filter_db = nil,
  -- history_dir = vim.fn.stdpath('data') .. '/teradata/history',
  -- bookmarks_dir = vim.fn.stdpath('data') .. '/teradata/bookmarks',
  -- data_dir = vim.fn.stdpath('data') .. '/teradata',
})

users (table[])

Default: {}

List of connection profiles. The first profile is the default active user.

users = {
  {
    log_mech = 'TD2',   -- logon mechanism (e.g. 'TD2', 'LDAP', 'KRB5')
    user     = 'USER',  -- username backed by tdwallet
    tdpid    = 'HOST',  -- Teradata host or IP address
  },
}

Security Note: Passwords are NEVER stored by vim-teradata or written to disk. They are retrieved at logon time from tdwallet using $tdwallet(<user>).

replacements (table)

Default: {}

String substitutions applied to every query before it is sent to BTEQ. Useful for portable scripts and environment-specific database names:

replacements = {
  ['${DB_NAME}'] = 'PROD_DB',
  ['${WORK_DB}'] = 'WORK_DB_001',
}

retlimit (number)

Default: 100

Maximum number of rows returned per :TDO / :TDV query. Controls memory and latency.

retlimit = 100

ft (string[])

Default: { "sql", "teradata" }

File extensions and filetypes that activate vim-teradata. The plugin registers each extension via vim.filetype.add. See Filetypes.

ft = { 'sql', 'teradata', 'btq' }

picker (string)

Default: "auto"

Picker backend used by :TDR and selection dialogs:

picker = 'auto'

sql_detection (string)

Default: "always"

Controls whether .sql files are treated as teradata:

  • "always": all .sql buffers get filetype teradata.
  • "never": do not map .sql to teradata.
  • "modeline": inspect first/last 5 lines for teradata or bteq.
sql_detection = 'always'

timeout_ms (number | nil)

Default: nil (disabled)

BTEQ job execution timeout in milliseconds. Set to nil or 0 to disable the timeout.

timeout_ms = 30000 -- 30 seconds

history_max (number | nil)

Default: nil (unlimited)

Maximum number of query history runs kept on disk before pruning older entries. Set to nil to disable pruning.

history_max = 500

keymaps (table)

Default: see below

Normal-mode mappings registered in every Teradata buffer to invoke node-aware editing functions ("OPE operations"). Each entry maps a key sequence to either:

  • A string: name of an OPE function applied at the cursor node.
  • A table { "func_name", { "node_type", ... } }: function plus the tree-sitter node type to act on around the cursor.

Default keymaps:

keymaps = {
  ["gcu"] = "uncomment_node",
  ["gcq"] = { "comment_node", { "statement" } },
  ["gcs"] = { "comment_node", { "select_expression" } },
  ["gcw"] = { "comment_node", { "where" } },
  ["gcf"] = { "comment_node", { "term" } },
  ["gcb"] = { "comment_node", { "binary_expression" } },
  ["gdq"] = { "delete_node",  { "statement" } },
  ["gds"] = { "delete_node",  { "select_expression" } },
  ["gdw"] = { "delete_node",  { "where" } },
  ["gdf"] = { "delete_node",  { "term" } },
  ["gdb"] = { "delete_node",  { "binary_expression" } },
  ["grs"] = { "copy_node",    { "select_expression" } },
  ["]s"]  = { "jump_to_next", { "statement" } },
  ["[s"]  = { "jump_to_prev", { "statement" } },
}

See Editing & Formatting for the full list of available OPE functions and node types.

sep (string)

Default: "~"

Field separator used in BTEQ result CSV files. Change only if your data regularly contains ~.

sep = '~'

edit_enabled (boolean)

Default: true

Enables in-grid editing of query results. See Editing results.

edit_enabled = true

preview_updates (boolean)

Default: true

When true, saving edits in a result buffer opens the generated UPDATE statements in a scratch buffer for review; nothing is sent to BTEQ until you press <CR> in that buffer.

preview_updates = true

null_token (string)

Default: "NULL"

Text BTEQ uses to render NULL in result exports (.set null). The same token typed into a cell in edit mode means SQL NULL.

null_token = 'NULL'

edit_keymaps (table)

Default: { toggle = "E", save = "S", cancel = "C" }

Result-buffer local mappings for in-grid editing mode:

edit_keymaps = {
  toggle = 'E',  -- toggle edit mode
  save   = 'S',  -- stage / preview updates
  cancel = 'C',  -- discard pending edits
}

rerun_keymaps (table)

Default: { sql = "g.", result = "." }

Mappings that re-run the latest query (see :TDRerun):

rerun_keymaps = {
  sql    = 'g.',  -- installed in every Teradata buffer
  result = '.',   -- installed in every result buffer
}

The two defaults differ on purpose: in a Teradata buffer . must keep its normal meaning (repeat the last change), so the mnemonic g. is used instead. A result buffer is nofile and non-modifiable, so . is free there.

Set either entry to nil or "" to skip that mapping:

rerun_keymaps = { sql = '<leader>rr', result = '.' }

filter_db (string | nil)

Default: nil

Optional Lua pattern that limits which database names are imported by :TDSync. Useful on systems with thousands of irrelevant catalogs.

filter_db = '^PROD_'  -- only import databases starting with PROD_

Directory paths

  • history_dir (string): Directory for query history files (queries/ and resultsets/). Defaults to vim.fn.stdpath('data') .. '/teradata/history'.
  • bookmarks_dir (string): Directory for global and user bookmarks. Defaults to vim.fn.stdpath('data') .. '/teradata/bookmarks'.
  • data_dir (string): Directory for the autocomplete metadata cache (sql-autocomplete). Defaults to vim.fn.stdpath('data') .. '/teradata'.