Exercise 13.3.3 - new code with {dm} package since {datamodelr} is no longer actively maintained

[AnttiRask]

Brief summary: Just an update, since the datamodelr package is no longer maintained, here are the same two exercises, but with {dm} syntax.

Link to page, and reference the exercise number (if applicable): https://jrnold.github.io/r4ds-exercise-solutions/relational-data.html#exercise-13.3.3

New code:

First exercise:

# dm1

# Data frames
Lahman::Batting
Lahman::People
Lahman::Salaries


# Adding tables
lahman_dm_no_keys <- dm(
  Batting  = Lahman::Batting,
  People   = Lahman::People,
  Salaries = Lahman::Salaries
)

lahman_dm_no_keys


# A dm is a list
names(lahman_dm_no_keys)

lahman_dm_no_keys$Batting

lahman_dm_no_keys[c("People", "Salaries")]


# Defining Keys

## Primary Keys - candidates
dm_enum_pk_candidates(
  dm    = lahman_dm_no_keys, 
  table = Batting
)

dm_enum_pk_candidates(
  dm    = lahman_dm_no_keys,
  table = People
)

dm_enum_pk_candidates(
  dm    = lahman_dm_no_keys,
  table = Salaries
)


## Primary Keys - add the keys
lahman_dm_only_pks <- lahman_dm_no_keys %>%
  dm_add_pk(table = Batting, columns = c(playerID, yearID, stint)) %>%
  dm_add_pk(People, playerID) %>%
  dm_add_pk(Salaries, c(yearID, teamID, playerID))

lahman_dm_only_pks


## Foreign Keys - candidates
dm_enum_fk_candidates(
  dm        = lahman_dm_only_pks,
  table     = Batting,
  ref_table = People
)

dm_enum_fk_candidates(
  dm        = lahman_dm_only_pks,
  table     = Salaries,
  ref_table = People
)


## Foreign Keys - add the keys
lahman_dm_all_keys <-
  
  lahman_dm_only_pks %>%
  
  dm_add_fk(table = Batting, columns = playerID, ref_table = People) %>%
  dm_add_fk(table = Salaries, columns = playerID, ref_table = People)

lahman_dm_all_keys


# Visualization

## No keys
lahman_dm_no_keys %>%
  dm_draw(
    rankdir = "TB",
    view_type = "all"
  )

## With a primary key
lahman_dm_no_keys %>%
  dm_add_pk(Batting, c(playerID, yearID, stint)) %>%
  dm_draw()

## With all the primary keys and a foreign key
lahman_dm_only_pks %>%
  dm_add_fk(Salaries, playerID, People) %>%
  dm_draw()

## With all the keys
lahman_dm_all_keys %>%
  dm_draw()

## With all the keys and all the other columns as well
lahman_dm_all_keys %>%
  dm_draw(rankdir = "LR", view_type = "all", column_types = TRUE)


# Integrity Checks
lahman_dm_no_keys %>%
  dm_examine_constraints()

lahman_dm_only_pks %>%
  dm_examine_constraints()

lahman_dm_all_keys %>%
  dm_examine_constraints()


# Programming

# Return a data frame with our primary keys
lahman_dm_only_pks %>% 
  dm_get_all_pks()

# Return a data frame with our foreign keys
lahman_dm_all_keys %>% 
  dm_get_all_fks()

# Using as_tibble() to programmatically inspect which constraints (if any) are not satisfied
lahman_dm_all_keys %>%
  dm_examine_constraints() %>%
  tibble::as_tibble()

Second exercise:

# dm2

# Data frames
Lahman::People
Lahman::Managers
Lahman::AwardsManagers


# Adding tables
lahman_dm2_no_keys <- dm(
  People         = Lahman::People,
  Managers       = Lahman::Managers,
  AwardsManagers = Lahman::AwardsManagers
)

lahman_dm2_no_keys


# A dm is a list
names(lahman_dm2_no_keys)

lahman_dm2_no_keys$People

lahman_dm2_no_keys[c("Managers", "AwardsManagers")]


# Defining Keys

## Primary Keys - candidates
dm_enum_pk_candidates(
  dm    = lahman_dm2_no_keys, 
  table = People
)

dm_enum_pk_candidates(
  dm    = lahman_dm2_no_keys,
  table = Managers
)

dm_enum_pk_candidates(
  dm    = lahman_dm2_no_keys,
  table = AwardsManagers
)


## Primary Keys - add the keys
lahman_dm2_only_pks <- lahman_dm2_no_keys %>%
  dm_add_pk(table = People, columns = playerID) %>%
  dm_add_pk(Managers, c(yearID, teamID, inseason)) %>%
  dm_add_pk(AwardsManagers, c(playerID, awardID, yearID))

lahman_dm2_only_pks


## Foreign Keys - candidates
dm_enum_fk_candidates(
  dm        = lahman_dm2_only_pks,
  table     = Managers,
  ref_table = People
)

dm_enum_fk_candidates(
  dm        = lahman_dm2_only_pks,
  table     = AwardsManagers,
  ref_table = People
)


## Foreign Keys - add the keys
lahman_dm2_all_keys <-
  
  lahman_dm2_only_pks %>%
  
  dm_add_fk(table = Managers, columns = playerID, ref_table = People) %>%
  dm_add_fk(table = AwardsManagers, columns = playerID, ref_table = People)

lahman_dm2_all_keys


# Visualization

## No keys
lahman_dm2_no_keys %>%
  dm_draw(
    rankdir = "TB",
    view_type = "all"
  )

## With a primary key
lahman_dm2_no_keys %>%
  dm_add_pk(Managers, c(yearID, teamID, inseason)) %>%
  dm_draw()

## With all the primary keys and a foreign key
lahman_dm2_only_pks %>%
  dm_add_fk(AwardsManagers, playerID, People) %>%
  dm_draw()

## With all the keys
lahman_dm2_all_keys %>%
  dm_draw()

## With all the keys and all the other columns as well
lahman_dm2_all_keys %>%
  dm_draw(rankdir = "LR", view_type = "all", column_types = TRUE)


# Integrity Checks
lahman_dm2_no_keys %>%
  dm_examine_constraints()

lahman_dm2_only_pks %>%
  dm_examine_constraints()

lahman_dm2_all_keys %>%
  dm_examine_constraints()


# Programming

# Return a data frame with our primary keys
lahman_dm2_only_pks %>% 
  dm_get_all_pks()

# Return a data frame with our foreign keys
lahman_dm2_all_keys %>% 
  dm_get_all_fks()

# Using as_tibble() to programmatically inspect which constraints (if any) are not satisfied
lahman_dm2_all_keys %>%
  dm_examine_constraints() %>%
  tibble::as_tibble()