---
title: "Apply and Dates, Updated using Map"
author: "Prof. Eric A. Suess"
date: "09/30/2020"
output:
  pdf_document: default
  html_notebook: default
  word_document: default
---

# Apply and Dates

Some examples from Chapter 5

### apply, updated using map() and map_df()

With the map functions there is no need for the MARGIN option to specify rows (1) or columns (2).

```{r message = FALSE}
library(pacman)
p_load(tidyverse, purr, stringr, Lahman)

Teams %>% select(15:40) %>%
  apply(MARGIN = 2, FUN = mean, na.rm = TRUE)

Teams %>% select(15:40) %>%
  map(., mean, na.rm=TRUE) %>%
  unlist()

Teams %>% select(15:40) %>%
  map_df(., mean, na.rm=TRUE)
```

### lapply and sapply

Note that lapply() returns a list.  The map()

```{r}
angles <- Teams %>%
  filter(franchID == "ANA") %>%
  group_by(teamID, name) %>%
  summarize(began = first(yearID), ended = last(yearID)) %>%
  arrange(began)
angles
```

```{r}
angles_names <- angles %>%
  ungroup(teamID) %>%
  select(name)
class(angles_names)

nchar(angles_names[1,1])
nchar(angles_names[2,1])
nchar(angles_names[3,1])
nchar(angles_names[4,1])
```

```{r}
x <- lapply(angles_names, FUN = nchar)
class(x)
x
```

```{r}
y <- sapply(angles_names, FUN = nchar)
class(y)
y
```
```{r}
z <- angles_names %>% map(., str_length)
class(z)
z
```

```{r}
z <- angles_names %>% map_df(., str_length)
class(z)
z
```

### Example of a function used with lapply

```{r}
top5 <- function(x, teamnames) {
  x %>% filter(name == teamnames) %>%
    select(teamID, yearID, W, L, name) %>%
    arrange(desc(W)) %>%
    head(n = 5)
}

angles_list <- lapply(angles_names, FUN = top5, x = Teams)
class(angles_list)
angles_list

angles_list <- angles_names %>% map(., top5, x = Teams)
class(angles_list)
angles_list
```

### Dates

```{r}
library(mdsr)

OrdwayBirds

OrdwayBirds %>% select(Timestamp, Year, Month, Day) %>%
  glimpse()
```

covert to numbers

```{r}
OrdwayBirds %>% select(Timestamp, Year, Month, Day) %>%
  glimpse()
```

convert Timestamp

```{r}
library(lubridate)

WhenAndWho <- OrdwayBirds %>%
  mutate(When = mdy_hms(Timestamp)) %>%
  select(Timestamp, Year, Month, Day, When, DataEntryPerson) %>%
  glimpse()
```

```{r}
WhenAndWho %>% ggplot(aes(x = When, y = DataEntryPerson)) +
  geom_point(alpha = 0.1, position = "jitter")
```


```{r}
WhenAndWho %>% group_by(DataEntryPerson) %>%
  summarize(start = first(When), finish = last(When)) %>%
  mutate( duration = interval(start, finish) / ddays(1) )
```


```{r}
now()

as.Date(now())

today()

as.Date(today())

as.Date(now()) - as.Date(today())
```


```{r}
now()

as_date(now())

today()

as_date(today())

as_date(now()) - as_date(today())
```

How many days have you woken up in the morning?  Change the date.

What is wrong with this?

```{r}
as.Date(today()) - as.Date("01/01/1970")
```
Note the use of the date format.

```{r}
as.Date(today()) - as.Date("01/01/1970", "%m/%d/%Y")
```
In the lubradate package there is the mdy() function.

```{r}
as_date(today()) 
as_date(mdy("01/01/1970"))
as_date(today()) - as_date(mdy("01/01/1970"))
```