---
title: "Connecting to Data"
vignette: >
  %\VignetteIndexEntry{connect_to_data}
  %\VignetteEngine{quarto::html}
  %\VignetteEncoding{UTF-8}
knitr:
  opts_chunk:
    collapse: true
    comment: '#>'
---



Mighty in itself does not read any data, but the generated programs will
need to connect to a data source to be able to execute. Thus, the programs
will use the
[`connector`](https://github.com/NovoNordisk-OpenSource/connector/tree/main?tab=readme-ov-file#readme)
package to connect to study data.

The `path_connector_config` argument to `generate_adam_code()` is the file
path to the connector config file, embedded verbatim into the generated
programs. Mighty does not read the file itself — it only embeds the path into
the generated code. The file is read later when the generated program runs.
The value can be a literal file path or an `!expr`-prefixed R expression for
runtime evaluation (e.g., `!expr here::here("_connector.yml")`).

An example of a generated program can be seen below, where the `cnt` object is

*   created by reading the connector config file at the path specified by
    `path_connector_config`,
*   used to read `DM` and `DM_VACCINE`, and finally
*   save the created `ADSL` data set.


```r
# LOAD all external datasets needed ------------------------------------------------

cnt <- connector::connect(config = "_connector.yml")

DM <- cnt$sdtm$read_cnt(tolower("DM")) |> dplyr::select(ACTARM, AGE, AGEU, ARM, COUNTRY, DOMAIN, DTHFL, RACE, SEX, STUDYID, USUBJID)

DM_VACCINE <- cnt$sdtm$read_cnt(tolower("DM_VACCINE")) |> dplyr::select(ACTARM, AGE, AGEU, ARM, COUNTRY, DOMAIN, DTHFL, RACE, SEX, STUDYID, USUBJID)

### Generated ADaM program code (removed for simplicity)

# Save ADSL ------------------------------------------------

cnt$adam$write_cnt(ADSL, "adsl.parquet", overwrite = TRUE)

# Remove input tables
rm(list = c("DM", "DM_VACCINE"))
```