19_functions_practice.knit

Writing Reusable Code with Functions

Let’s Say You Have This…

You Want To…

dat  <- make_buoy_data(1986)

ggplot(dat, 
       aes(x = Date, y = Temperature_c)) + 
  geom_line()

But There Are Problems - Changing Year Formats and names

# A tibble: 6 × 16
     YY    MM    DD    hh    WD  WSPD   GST  WVHT   DPD   APD   MWD   BAR  ATMP
  <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl>
1    86     1     1     0   210     9    11    99    99    99   999 1009.   5.8
2    86     1     1     1   210     9    10    99    99    99   999 1009.   4.7
3    86     1     1     2   220     8    10    99    99    99   999 1009.   4  
4    86     1     1     3   220     6     8    99    99    99   999 1009.   3.9
5    86     1     1     4   210     6     8    99    99    99   999 1009    3.9
6    86     1     1     5   210     6     8    99    99    99   999 1008.   3.9
# ℹ 3 more variables: WTMP <dbl>, DEWP <dbl>, VIS <dbl>

# A tibble: 6 × 16
   YYYY    MM    DD    hh    WD  WSPD   GST  WVHT   DPD   APD   MWD   BAR  ATMP
  <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl>
1  1999     1     1     0   221   5.4   7.2  0.33  11.1  5.25   999 1017.  -4.3
2  1999     1     1     1   218   5.6   7.3  0.31  11.1  5.51   999 1016.  -4  
3  1999     1     1     2   226   5.7   7    0.32  12.5  6.53   999 1016.  -3.9
4  1999     1     1     3   228   5.7   7.3  0.31  11.1  6.17   999 1015.  -3.9
5  1999     1     1     4   237   5.9   7.7  0.39  11.1  5.02   999 1015   -3.8
6  1999     1     1     5   235   5.7   7.5  0.42  11.1  4.81   999 1015.  -3.9
# ℹ 3 more variables: WTMP <dbl>, DEWP <dbl>, VIS <dbl>

Worse… nonnumeric row cruft

# A tibble: 6 × 18
  X.YY  MM    DD    hh    mm    WDIR  WSPD  GST   WVHT  DPD   APD   MWD   PRES  
  <chr> <chr> <chr> <chr> <chr> <chr> <chr> <chr> <chr> <chr> <chr> <chr> <chr> 
1 #yr   mo    dy    hr    mn    degT  m/s   m/s   m     sec   sec   degT  hPa   
2 2012  06    30    23    50    253   3.4   4.2   0.52  9.09  5.33  86    1005.4
3 2012  07    01    00    50    198   3.5   4.1   0.53  7.69  5.41  107   1005.7
4 2012  07    01    01    50    199   4.0   4.7   0.48  10.00 5.60  102   1006.1
5 2012  07    01    02    50    218   3.9   4.4   0.45  9.09  5.20  87    1006.1
6 2012  07    01    03    50    222   4.5   4.8   0.41  7.69  4.68  100   1006.5
# ℹ 5 more variables: ATMP <chr>, WTMP <chr>, DEWP <chr>, VIS <chr>, TIDE <chr>

Solutions?

Solutions

Write code for each individual file
- Lots of copy-paste
Write code that row-binds all data and clean from there
- Only good once - what about next year?
FUNCTIONS!

What is our Philosophy?

“You should consider writing a function whenever you’ve copied and pasted a block of code more than twice”

- H. Wickham

Let’s Find the Problems!

Work with a partner, load in data files individually, and identify all of the problems/things that are different from file to file. Or, things that might make it difficult to work with data later (i.e., hard to understand what is in the data).

Problems with Buoys - Each one (or two) is a function!

NA characters vary (99, 999).
Need to eliminate rows with #yr & turn cols numeric
Names for year are nonstandard
- X.YY, YY, and YYYY
Some years are listed with only two digits
- Typically, but maybe not always, less than 1999
Col names are abbreviations - make it readable!
Is hourly really a good idea?

Where to Start?!

Every problem is an opportunity for a function.
With every problem, identify what you want to input as an argument, and what you want as an output.
Don’t make one function do too much! You can always chain together many functions

To start - reading in a buoy file

Let’s just write a function that will read in any of the buoys and return a nice data frame.
Let’s call it read_buoy.
But before we start….

How do I build a function?

OK, what’s normally going to change
- The year!
BUT - what could change under some circumstances?
- buoy number, directory, separator.
- let’s leave that for later and just
Identify what could differ between files for just loading
- Don’t try and do too much at once!
Write some test code to load in one buoy NOT in a function to start

Test Code: Copy these comments and write code!

# define a year

# make a filename

# read in a file

Test Code Example

# define a year
a_year <- 2013

# make a filename (could use paste0 instead of str_c)
# or glue::glue()
buoy_file <- str_c("lectures/data/buoydata/44013_", a_year, ".csv")

# read in a file
read_csv(buoy_file)

# A tibble: 8,210 × 18
   X.YY  MM    DD    hh    mm    WDIR  WSPD  GST   WVHT  DPD   APD   MWD   PRES 
   <chr> <chr> <chr> <chr> <chr> <chr> <chr> <chr> <chr> <chr> <chr> <chr> <chr>
 1 #yr   mo    dy    hr    mn    degT  m/s   m/s   m     sec   sec   degT  hPa  
 2 2012  12    31    23    50    234   8.5   10.2  0.42  3.33  3.14  279   1011…
 3 2013  01    01    00    50    246   9.0   10.9  0.52  2.86  2.94  251   1010…
 4 2013  01    01    01    50    243   7.3   9.4   0.50  3.33  3.05  237   1010…
 5 2013  01    01    02    50    252   10.5  12.4  0.50  3.03  2.99  244   1009…
 6 2013  01    01    03    50    256   10.9  13.0  0.65  3.70  3.28  265   1008…
 7 2013  01    01    04    50    254   11.6  13.2  0.72  3.70  3.33  255   1007…
 8 2013  01    01    05    50    244   10.8  12.7  0.84  4.00  3.54  260   1006…
 9 2013  01    01    06    50    241   10.5  12.4  0.74  3.85  3.34  259   1006…
10 2013  01    01    07    50    237   9.7   11.9  0.67  3.70  3.17  257   1006…
# ℹ 8,200 more rows
# ℹ 5 more variables: ATMP <chr>, WTMP <chr>, DEWP <chr>, VIS <chr>, TIDE <chr>

Unit Testing

Take the smallest piece of code in an application, and make a test to ensure if works properly

#define a year
a_year <- 1993

#make a filename
buoy_file <- str_c("lectures/data/buoydata/44013_", a_year, ".csv")

# read in a file
dat <- read_csv(buoy_file)

# A TEST (or two)
visdat::vis_dat(dat)
summary(dat)

Unit test for 1993 - OK on type

Unit test for 1993 - bad on range!

       YY           MM               DD              hh             WD       
 Min.   :93   Min.   : 1.000   Min.   : 1.00   Min.   : 0.0   Min.   :  0.0  
 1st Qu.:93   1st Qu.: 4.000   1st Qu.: 8.00   1st Qu.: 5.0   1st Qu.:124.0  
 Median :93   Median : 7.000   Median :16.00   Median :11.0   Median :209.0  
 Mean   :93   Mean   : 6.538   Mean   :15.72   Mean   :11.5   Mean   :195.3  
 3rd Qu.:93   3rd Qu.:10.000   3rd Qu.:23.00   3rd Qu.:18.0   3rd Qu.:278.0  
 Max.   :93   Max.   :12.000   Max.   :31.00   Max.   :23.0   Max.   :999.0  
      WSPD             GST              WVHT             DPD        
 Min.   : 0.000   Min.   : 0.000   Min.   : 0.000   Min.   : 0.000  
 1st Qu.: 4.000   1st Qu.: 4.700   1st Qu.: 0.400   1st Qu.: 5.900  
 Median : 5.900   Median : 7.000   Median : 0.600   Median : 8.300  
 Mean   : 6.382   Mean   : 8.138   Mean   : 2.247   Mean   : 9.351  
 3rd Qu.: 8.200   3rd Qu.:10.000   3rd Qu.: 1.000   3rd Qu.:10.000  
 Max.   :99.000   Max.   :99.000   Max.   :99.000   Max.   :99.000  
      APD              MWD           BAR              ATMP       
 Min.   : 0.000   Min.   :  0   Min.   : 964.6   Min.   :-14.00  
 1st Qu.: 4.500   1st Qu.:999   1st Qu.:1010.7   1st Qu.:  2.80  
 Median : 5.600   Median :999   Median :1016.4   Median :  9.50  
 Mean   : 7.042   Mean   :792   Mean   :1035.0   Mean   : 10.58  
 3rd Qu.: 6.800   3rd Qu.:999   3rd Qu.:1022.5   3rd Qu.: 16.50  
 Max.   :99.000   Max.   :999   Max.   :9999.0   Max.   :999.00  
      WTMP             DEWP          VIS       
 Min.   :  0.80   Min.   :999   Min.   : 0.00  
 1st Qu.:  4.00   1st Qu.:999   1st Qu.:99.00  
 Median :  9.50   Median :999   Median :99.00  
 Mean   : 11.67   Mean   :999   Mean   :78.51  
 3rd Qu.: 14.10   3rd Qu.:999   3rd Qu.:99.00  
 Max.   :999.00   Max.   :999   Max.   :99.00

What are other problems these unit tests find? Try it out!

More testing with Small Changes - YES!

read_csv(buoy_file, na = c("99", "999", "9999")) |>
 summary()

       YY           MM               DD              hh             WD       
 Min.   :93   Min.   : 1.000   Min.   : 1.00   Min.   : 0.0   Min.   :  0.0  
 1st Qu.:93   1st Qu.: 4.000   1st Qu.: 8.00   1st Qu.: 5.0   1st Qu.:124.0  
 Median :93   Median : 7.000   Median :16.00   Median :11.0   Median :209.0  
 Mean   :93   Mean   : 6.538   Mean   :15.72   Mean   :11.5   Mean   :195.1  
 3rd Qu.:93   3rd Qu.:10.000   3rd Qu.:23.00   3rd Qu.:18.0   3rd Qu.:278.0  
 Max.   :93   Max.   :12.000   Max.   :31.00   Max.   :23.0   Max.   :360.0  
                                                              NA's   :20     
      WSPD            GST              WVHT             DPD        
 Min.   : 0.00   Min.   : 0.000   Min.   :0.0000   Min.   : 0.000  
 1st Qu.: 4.00   1st Qu.: 4.700   1st Qu.:0.4000   1st Qu.: 5.900  
 Median : 5.90   Median : 7.000   Median :0.6000   Median : 8.300  
 Mean   : 6.34   Mean   : 7.739   Mean   :0.8358   Mean   : 8.043  
 3rd Qu.: 8.20   3rd Qu.: 9.900   3rd Qu.:1.0000   3rd Qu.:10.000  
 Max.   :25.70   Max.   :32.400   Max.   :6.1000   Max.   :16.700  
 NA's   :4       NA's   :38       NA's   :125      NA's   :125     
      APD            MWD             BAR              ATMP        
 Min.   : 0.0   Min.   :  0.0   Min.   : 964.6   Min.   :-14.000  
 1st Qu.: 4.5   1st Qu.: 73.0   1st Qu.:1010.6   1st Qu.:  2.800  
 Median : 5.6   Median : 91.0   Median :1016.4   Median :  9.500  
 Mean   : 5.7   Mean   :117.6   Mean   :1016.4   Mean   :  9.211  
 3rd Qu.: 6.7   3rd Qu.:111.0   3rd Qu.:1022.4   3rd Qu.: 16.500  
 Max.   :11.8   Max.   :358.0   Max.   :1044.5   Max.   : 31.900  
 NA's   :125    NA's   :6677    NA's   :20       NA's   :12       
      WTMP          DEWP              VIS       
 Min.   : 0.800   Mode:logical   Min.   : 0.00  
 1st Qu.: 4.000   NA's:8693      1st Qu.: 7.30  
 Median : 9.400                  Median : 9.10  
 Mean   : 9.503                  Mean   :11.74  
 3rd Qu.:14.100                  3rd Qu.:11.00  
 Max.   :22.000                  Max.   :35.80  
 NA's   :19                      NA's   :6652

Tested against a few years to figure out NAs, vis_dat helped

Now ask, what more can be generalized for arguments?

Make this a Function by making a_year an argument and adding a return!

#define a year
a_year <- 2013

#make a filename
buoy_file <- str_c("lectures/data/buoydata/44013_", a_year, ".csv")

# read in a file
read_csv(buoy_file, na = c("99", "999", "9999"))

Also, get a little fancier and add a buoy_number and buoy_dir argument with sensible defaults - you’ve got this!

One Reading Function

read_buoy <- function(a_year, buoy_number = 44013, 
                      buoy_dir = "lectures/data/buoydata/"){

  #make a file name
  buoy_file <- str_c(buoy_dir, buoy_number, "_", a_year, ".csv")
  
  #read in the file
  one_buoy <- read_csv(buoy_file, 
                       na = c("99", "999", "9999"))
  
  #return the file
  return(one_buoy)
  
}

Problems with Buoys - Each one (or two) is a function!

~~NA characters vary (99, 999).~~
Need to eliminate rows with #yr & turn cols numeric
Names for year are nonstandard
- X.YY, YY, and YYYY
Some years are listed with only two digits
- Typically, but maybe not always, less than 1999
Col names are abbreviations - make it readable!
Is hourly really a good idea?

Problem 2: Extra Rows and Characters

# A tibble: 6 × 18
  X.YY  MM    DD    hh    mm    WDIR  WSPD  GST   WVHT  DPD   APD   MWD   PRES  
  <chr> <chr> <chr> <chr> <chr> <chr> <chr> <chr> <chr> <chr> <chr> <chr> <chr> 
1 #yr   mo    dy    hr    mn    degT  m/s   m/s   m     sec   sec   degT  hPa   
2 2012  12    31    23    50    234   8.5   10.2  0.42  3.33  3.14  279   1011.6
3 2013  01    01    00    50    246   9.0   10.9  0.52  2.86  2.94  251   1010.8
4 2013  01    01    01    50    243   7.3   9.4   0.50  3.33  3.05  237   1010.0
5 2013  01    01    02    50    252   10.5  12.4  0.50  3.03  2.99  244   1009.0
6 2013  01    01    03    50    256   10.9  13.0  0.65  3.70  3.28  265   1008.1
# ℹ 5 more variables: ATMP <chr>, WTMP <chr>, DEWP <chr>, VIS <chr>, TIDE <chr>

Fix it with a function!

Work with a partner and write a function that will strip non-numeric rows from a buoy data frame and turn character cols numeric
Start with a single test case, and apply unit tests.
- hint: mutate(across(.cols = everything(), .fns = as.numeric))
- hint: what is the same and what is different across files?
Once it works broadly, turn it into a function so that we can…

read_buoy(1995) |>
  fix_bad_buoy_rows()

My test code outline (there are many solutions)

# get some buoy data

# Turn everything numeric, as then some rows will be all NA

# Remove rows where month is NA, as the year col name changes

# TESTS

Alternate Solution if we Don’t Care About NAs

# get some buoy data

# Turn everything numeric, as then some rows will be all NA

# TESTS

My test code

# get some buoy data
dat <- read_buoy(2013)

# Turn everything numeric, as then some rows will be all NA
dat <- dat |>
  mutate(across(.cols = everything(),
                .fns = as.numeric))

# Remove rows where month is NA, as the year col name changes
dat <- dat |> 
  filter(!is.na(MM))

# TESTS
visdat::vis_dat(dat)
summary(dat)

Turn it into a function

fix_bad_buoy_rows <- function(dat) {
  
  # Turn everything numeric, as then some rows will be all NA
  dat <- dat |>
    mutate(across(.cols = everything(),
                  .fns = as.numeric)) |>
  
  # Remove rows where month is NA, as the year col name changes
    filter(!is.na(MM))
  
  # return
  return(dat)
}

Does It Blend?

read_buoy(2013) |>
  fix_bad_buoy_rows() |>
  head()

# A tibble: 6 × 18
   X.YY    MM    DD    hh    mm  WDIR  WSPD   GST  WVHT   DPD   APD   MWD  PRES
  <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl>
1  2012    12    31    23    50   234   8.5  10.2  0.42  3.33  3.14   279 1012.
2  2013     1     1     0    50   246   9    10.9  0.52  2.86  2.94   251 1011.
3  2013     1     1     1    50   243   7.3   9.4  0.5   3.33  3.05   237 1010 
4  2013     1     1     2    50   252  10.5  12.4  0.5   3.03  2.99   244 1009 
5  2013     1     1     3    50   256  10.9  13    0.65  3.7   3.28   265 1008.
6  2013     1     1     4    50   254  11.6  13.2  0.72  3.7   3.33   255 1007.
# ℹ 5 more variables: ATMP <dbl>, WTMP <dbl>, DEWP <dbl>, VIS <dbl>, TIDE <dbl>

Problems with Buoys - Each one (or two) is a function!

~~NA characters vary (99, 999).~~
~~Need to eliminate rows with #yr & turn cols numeric~~
Names for year are nonstandard
- X.YY, YY, and YYYY
Some years are listed with only two digits
- Typically, but maybe not always, less than 1999
Col names are abbreviations - make it readable!
Is hourly really a good idea?

That Year Problem

# A tibble: 6 × 16
     YY    MM    DD    hh    WD  WSPD   GST  WVHT   DPD   APD   MWD   BAR  ATMP
  <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl>
1    86     1     1     0   210     9    11    NA    NA    NA    NA 1009.   5.8
2    86     1     1     1   210     9    10    NA    NA    NA    NA 1009.   4.7
3    86     1     1     2   220     8    10    NA    NA    NA    NA 1009.   4  
4    86     1     1     3   220     6     8    NA    NA    NA    NA 1009.   3.9
5    86     1     1     4   210     6     8    NA    NA    NA    NA 1009    3.9
6    86     1     1     5   210     6     8    NA    NA    NA    NA 1008.   3.9
# ℹ 3 more variables: WTMP <dbl>, DEWP <dbl>, VIS <dbl>

# A tibble: 6 × 16
   YYYY    MM    DD    hh    WD  WSPD   GST  WVHT   DPD   APD   MWD   BAR  ATMP
  <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl>
1  1999     1     1     0   221   5.4   7.2  0.33  11.1  5.25    NA 1017.  -4.3
2  1999     1     1     1   218   5.6   7.3  0.31  11.1  5.51    NA 1016.  -4  
3  1999     1     1     2   226   5.7   7    0.32  12.5  6.53    NA 1016.  -3.9
4  1999     1     1     3   228   5.7   7.3  0.31  11.1  6.17    NA 1015.  -3.9
5  1999     1     1     4   237   5.9   7.7  0.39  11.1  5.02    NA 1015   -3.8
6  1999     1     1     5   235   5.7   7.5  0.42  11.1  4.81    NA 1015.  -3.9
# ℹ 3 more variables: WTMP <dbl>, DEWP <dbl>, VIS <dbl>

That Year Problem

# A tibble: 6 × 16
     YY    MM    DD    hh    WD  WSPD   GST  WVHT   DPD   APD   MWD   BAR  ATMP
  <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl>
1    86     1     1     0   210     9    11    NA    NA    NA    NA 1009.   5.8
2    86     1     1     1   210     9    10    NA    NA    NA    NA 1009.   4.7
3    86     1     1     2   220     8    10    NA    NA    NA    NA 1009.   4  
4    86     1     1     3   220     6     8    NA    NA    NA    NA 1009.   3.9
5    86     1     1     4   210     6     8    NA    NA    NA    NA 1009    3.9
6    86     1     1     5   210     6     8    NA    NA    NA    NA 1008.   3.9
# ℹ 3 more variables: WTMP <dbl>, DEWP <dbl>, VIS <dbl>

# A tibble: 6 × 18
   X.YY    MM    DD    hh    mm  WDIR  WSPD   GST  WVHT   DPD   APD   MWD  PRES
  <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl>
1  2012     6    30    23    50   253   3.4   4.2  0.52  9.09  5.33    86 1005.
2  2012     7     1     0    50   198   3.5   4.1  0.53  7.69  5.41   107 1006.
3  2012     7     1     1    50   199   4     4.7  0.48 10     5.6    102 1006.
4  2012     7     1     2    50   218   3.9   4.4  0.45  9.09  5.2     87 1006.
5  2012     7     1     3    50   222   4.5   4.8  0.41  7.69  4.68   100 1006.
6  2012     7     1     4    50   231   3.4   3.8  0.4   9.09  5.24    82 1006.
# ℹ 5 more variables: ATMP <dbl>, WTMP <dbl>, DEWP <dbl>, VIS <dbl>, TIDE <dbl>

Fix two problems with 1 function!

Work with a partner and write a function that will:
- make the year column always be YYYY
- make the year column be a 4 digit year (e.g. 98 = 1998)
- hint: this is a good order to solve in
- hint: regexp with names() or rename_with() are helpful
Test & once it works broadly, turn it into a function so that we can…

read_buoy(1995) |>
  fix_bad_buoy_rows() |>
  standardize_buoy_years()

My test code

# get buoy data
dat <- read_buoy(1986)|> fix_bad_buoy_rows()

# use rename_with to fix year column name
dat <- dat |> rename_with(
  .cols = contains("YY"),
  .fn = ~"YYYY"
)

# if a year is < 100, add 1900 to it
dat <- dat |> 
  mutate(YYYY = ifelse(YYYY < 100, YYYY+1900, YYYY))

# TEST
range(dat$YYYY)

My alternate code

# get buoy data
dat <- read_buoy(1986)|> fix_bad_buoy_rows()

# use regexp to fix year column name
names(dat) <- str_replace( names(dat),
                           pattern = "X\\.YY", 
                           replace = "YYYY")

names(dat) <- str_replace( names(dat),
                           pattern = "^YY$", 
                           replace = "YYYY")


# if a year is < 100, add 1900 to it
dat <- dat |> 
  mutate(YYYY = ifelse(YYYY < 100, YYYY+1900, YYYY))

# TEST
range(dat$YYYY)

Make it into a function

standardize_buoy_years <- function(dat) {
  
  # use rename_with to fix year column name
  dat <- dat |> rename_with(.cols = contains("YY"),
                            .fn = ~ "YYYY") |>
    
    # if a year is < 100, add 1900 to it
    mutate(YYYY = ifelse(YYYY < 100, YYYY + 1900, YYYY))
  
  return(dat)
}

Does It Blend?

read_buoy(1995) |>
  fix_bad_buoy_rows() |>
  standardize_buoy_years() |> 
  head()

Does It Blend?

read_buoy(2013) |>
  fix_bad_buoy_rows() |>
  standardize_buoy_years() |> 
  head()

Problems with Buoys - Each one (or two) is a function!

~~NA characters vary (99, 999).~~
~~Need to eliminate rows with #yr & turn cols numeric~~
~~Names for year are nonstandard~~
- ~~X.YY, YY, and YYYY~~
~~Some years are listed with only two digits~~
- ~~Typically, but maybe not always, less than 1999~~
Col names are abbreviations - make it readable!
Is hourly really a good idea?

Before We Move Forward…

What is this?

read_buoy(1995) |>
  fix_bad_buoy_rows() |>
  standardize_buoy_years()

Note that you have been making little function MODULES that we string together

Year can still change - it’s almost like we have another function in the making composed of modules…. this is….

Modular Programming

Wait, what is this modular madness?

Our code is composed of modular pieces
This means each modular function can be used in other contexts
- Enhances code flexibility and future reuse
It also means it is easier to find where our code has gone wrong
- Easier to test and debug
AND - we don’t get lost writing one honking single function

From Pipelines to Wrapper Function

read_buoy(1995) |>
  fix_bad_buoy_rows() |>
  standardize_buoy_years()

Let’s wrap this in a function

get_buoy_data <- function(a_year){
  read_buoy(a_year) |>
  fix_bad_buoy_rows() |>
  standardize_buoy_years() 
}

A Cleaner Wrapper

get_buoy_data <- function(a_year){
  
  # read in buoy file and fix NAs
  read_buoy(a_year) |>
  
  # fix non-numeric row problem
  fix_bad_buoy_rows() |>
  
  # standardize year column
  standardize_buoy_years() 
}

Backwards Approach to Writing Modular Code

Analyze the steps of your task.
Break it down into pieces.
Write one function per piece.
Chain pieces together in a pipeline.
Turn pipeline into a wrapper function!

Or - Thinking Forwards for Writing Modular Code

Start with a wrapper
- Write out the steps of what you want to do in comments
- Pair those with functions using “snake case” clean_data
- Use |> for efficiency/readability!
Write the subfunctions
- Repeat the toplevel comments anf functions
Continue until you have the lowest level atomized functions
Test THOSE functions. Then test up the chain….

Why do this

Reusable modules in other context.
Can unit test elements of the wrapper function
Easier to debug!
Can solve problems forwards or backwards!

Reality Check

Yes, sometimes you will write long functions - as long as they are simple and readable, meh

Final Exercises - do 1, and then make wrappers in 2 and 3 with MINIMAL effort

Write a function that will get daily summaries of wave height and temperature only
- Make the returned column names for all values nice and readable
- use paste() and as.Date() to make a Date column after summarize
Do the same, but monthly…
Do the same, but anually….