remove dbplyr

Author: cmaimone

r/exercises_with_answers.Rmd

@@ -91,12 +91,11 @@ dbDisconnect(con)
 ```
 
 
-## Exercise: `dbplyr`
+## Exercise: `dplyr`
 
-Connect to the dvdrental database.  Repeat [Exercise: Joining and Grouping 2](https://github.com/nuitrcs/databases_workshop/blob/master/sql/part2_exercises_with_answers.md) from Part 2 using `dbplyr`. 
+Connect to the dvdrental database.  Repeat [Exercise: Joining and Grouping 2](https://github.com/nuitrcs/databases_workshop/blob/master/sql/part2_exercises_with_answers.md) from Part 2 using `dplyr`. 
 
 ```{r, echo=TRUE, eval=TRUE}
-library(dbplyr)
 library(dplyr)
 ```
 

r/exercises_with_answers.html

@@ -11,7 +11,7 @@
 
 <meta name="author" content="Christina Maimone" />
 
-<meta name="date" content="2019-08-09" />
+<meta name="date" content="2019-08-12" />
 
 <title>R Database Exercises</title>
 
@@ -377,7 +377,7 @@
 
 <h1 class="title toc-ignore">R Database Exercises</h1>
 <h4 class="author">Christina Maimone</h4>
-<h4 class="date">2019-08-09</h4>
+<h4 class="date">2019-08-12</h4>
 
 </div>
 
@@ -425,19 +425,18 @@ <h4>Solution</h4>
 }
 selectedrows</code></pre>
 <pre><code>##    id  name
-## 1 269 ZNUET
-## 2 475 VPMBT
-## 3 409 AZIQX</code></pre>
+## 1  99 UIBJN
+## 2 236 YLHQP
+## 3 596 BYEAM</code></pre>
 <p>We used paste function above because we have control over offset – it would be better to use a prepared query, but since we aren’t getting input from a user, it’s not super dangerous.</p>
 <p>An alternative approach, which could work well if the table isn’t too big, is to retrieve all of the IDs, and then randomly sample the IDs, and retrieve just those rows.</p>
 <pre class="r"><code>dbDisconnect(con)</code></pre>
 </div>
 </div>
-<div id="exercise-dbplyr" class="section level2">
-<h2>Exercise: <code>dbplyr</code></h2>
-<p>Connect to the dvdrental database. Repeat <a href="https://github.com/nuitrcs/databases_workshop/blob/master/sql/part2_exercises_with_answers.md">Exercise: Joining and Grouping 2</a> from Part 2 using <code>dbplyr</code>.</p>
-<pre class="r"><code>library(dbplyr)
-library(dplyr)</code></pre>
+<div id="exercise-dplyr" class="section level2">
+<h2>Exercise: <code>dplyr</code></h2>
+<p>Connect to the dvdrental database. Repeat <a href="https://github.com/nuitrcs/databases_workshop/blob/master/sql/part2_exercises_with_answers.md">Exercise: Joining and Grouping 2</a> from Part 2 using <code>dplyr</code>.</p>
+<pre class="r"><code>library(dplyr)</code></pre>
 <div id="solution-1" class="section level4">
 <h4>Solution</h4>
 <p>Set your connection information as appropriate for the workshop:</p>

r/r_databases.Rmd

@@ -35,7 +35,7 @@ library(RPostgres)
 
 We connect with a function call like the following.
 
-Note: this code was generated on my local machine connected to a local copy of the database.
+Note: this code was generated on my local machine connected to a local copy of the database.  Your connection details will be different.  Note I also have permissions to modify this database.
 
 ```{r}
 con <- dbConnect(RPostgres::Postgres(), host="localhost", dbname="dvdrental")
@@ -88,7 +88,7 @@ If you want part of your query to be determined by a variable -- especially if i
 ```{r}
 # YES
 myquery <- dbSendQuery(con, "select * from actor where actor_id = $1")
-dbBind(myquery, list(5))
+dbBind(myquery, list(4))
 dbFetch(myquery)
 ```
 
@@ -124,8 +124,12 @@ Which are ok, but could get annoying.
 
 If you're not a superuser on the `dvdrental` database, just try connecting to a database you can modify. Then the basic function is `dbSendQuery` for any command you want to execute where you aren't retrieving results.
 
+Note that by default, statements take effect immediately - they are not in a transaction that you need to commit.  To use transactions, see below.
+
 ```{r, eval=FALSE}
-dbSendQuery(con, statement="update actor set actor_id=5000 where actor_id=5")
+res <- dbSendQuery(con, statement="update actor set first_name='Jenn' where actor_id=4")
+print(res)  # contains info on result of update
+dbClearResult(res)  # prevent warning messages
 ```
 
 To create a table, you can give it a data frame
@@ -144,6 +148,28 @@ dbRemoveTable(con, "mynewtable")
 ```
 
 
+## Transactions
+
+There are also methods for managing transactions if you need: `dbBegin`, `dbRollback`, `dbCommit`. Transactions are key for when you need to be sure that a sequence of SQL commands (e.g. `UPDATE`, `CREATE`, `DROP`, `DELETE`, etc.) execute correctly before they're made permanent (i.e. "committed").
+
+
+```{r, eval=FALSE}
+dbBegin(con)
+dbWriteTable(con, "mynewtable", mytbl)
+dbRollback(con)
+dbGetQuery(con, "SELECT * FROM mynewtable")
+```
+
+The above will produce error:
+
+```
+Error in result_create(conn@ptr, statement) : 
+  Failed to prepare query: ERROR:  relation "mynewtable" does not exist
+LINE 1: SELECT * FROM mynewtable
+```
+
+because the transaction was rolled back, not committed.
+
 ## Close Connection
 
 Connections will get closed when you quit R, but it's good practice to explicitly close them.
@@ -152,17 +178,15 @@ Connections will get closed when you quit R, but it's good practice to explicitl
 dbDisconnect(con)
 ```
 
-## Transactions
 
-There are also methods for managing transactions if you need: `dbBegin`, `dbRollback`, `dbCommit`. Transactions are key for when you need to be sure that a sequence of SQL commands (e.g. `UPDATE`, `CREATE`, `DROP`, `DELETE`, etc.) execute correctly before they're made permanent (i.e. "committed").
 
 
 # Use `dplyr`
 
 For more complete info, see the [RStudio databases site](http://db.rstudio.com/dplyr/).
 
 ```{r, eval=FALSE}
-needToInstall <- c("dbplyr", "tidyverse")
+needToInstall <- c("tidyverse")
 needToInstall <- needToInstall[which(!needToInstall %in% installed.packages())]
 if(length(needToInstall) > 0){
   sapply(needToInstall, install.packages)
@@ -172,7 +196,6 @@ if(length(needToInstall) > 0){
 
 ```{r, message=FALSE, warning=FALSE}
 library(tidyverse)
-library(dbplyr)
 ```
 
 First, connect like normal
@@ -193,7 +216,7 @@ If we look at this object, it doesn't have data in it:
 str(actortbl)
 ```
 
-It just has connection information.  `dbplyr` will try to perform operations within the database where it can, instead of pulling all of the data into R.
+It just has connection information.  `dplyr` will try to perform operations within the database where it can, instead of pulling all of the data into R.
 
 Yet you can print the object and see observations:
 
@@ -229,7 +252,7 @@ rentaltbl %>%
   show_query()
 ```
 
-You can use `collect` to pull down all of the data (tell `dbplyr` to stop being lazy).
+You can use `collect` to pull down all of the data (tell `dplyr` to stop being lazy).
 
 ```{r, echo=TRUE}
 # First, without collecting
@@ -242,14 +265,20 @@ df1
 Looks OK, except:
 
 ```{r, eval=FALSE}
-df[1,]
+df1[1,]
 ```
 
 Gives you:
 
-`Error in df[1, ] : object of type 'closure' is not subsettable`
+`Error in df1[1, ] : incorrect number of dimensions`
+
+It's the wrong dimensions because `df1` isn't actually a data.frame:
+
+```{r}
+str(df1)
+```
 
-Which is a strange error, but it is telling us we need to collect the data.
+It is telling us we need to collect the data first to actually pull it into R.
 
 ```{r, echo=TRUE}
 # Then with collecting
@@ -276,14 +305,14 @@ custtbl %>%
 ```
 
 
-You could create a table with `copy_to` (if you have write permissions)
+You could create a table with `copy_to` (if you have the correct permissions)
 
 ```{r, scho=TRUE, eval=FALSE}
 mytbl <-data.frame(number=1:10 , letter=LETTERS[1:10])
 copy_to(con, mytbl, "mynewtable")
 ```
 
-By default, it creates a temporary table.  But this is a setting you can change, and you can also specify what columns to index on the table.  
+By default, it creates a **temporary** table.  But this is a setting you can change, and you can also specify what columns to index on the table.  
 
 
 Disconnect like we normally do