Package 'crplyr'

Flatten a Crunch Cube

Description

Crunch Cubes can be expressed as a long data frame instead of a multidimensional array. In this form each dimension of the cube is a variable and the cube values are expressed as columns for each measure. This is useful both to better understand what each entry of a cube represents, and to work with the cube result using tidyverse tools.

Usage

as_cr_tibble(x, ...)

Arguments

x	a CrunchCube
...	further arguments passed on to tibble::as_tibble()

Details

The cr_tibble class is a subclass of tibble that has extra metadata to allow ggplot::autoplot() to work. If you find that this extra metadata is getting in the way, you can use as_tibble() to get a true tibble.

---

Autoplot methods for Crunch Objects

Description

The Crunch autoplot methods generate ggplots that are tailored to various Crunch objects. This allows you to visualize the object without bringing it into memory. You can select between three families of plots, which will attempt to accomodate the dimensionality of the plotted object. These plots can be further extended and customized with other ggplot methods.

Usage

## S3 method for class 'DatetimeVariable'
autoplot(object, ...)

## S3 method for class 'NumericVariable'
autoplot(object, ...)

## S3 method for class 'CategoricalVariable'
autoplot(object, ...)

## S3 method for class 'CategoricalArrayVariable'
autoplot(object, ...)

## S3 method for class 'MultipleResponseVariable'
autoplot(object, ...)

## S3 method for class 'CrunchCube'
autoplot(object, ...)

## S3 method for class 'CrunchCubeCalculation'
autoplot(object, plot_type = "dot", ...)

## S3 method for class 'tbl_crunch_cube'
autoplot(object, plot_type = c("dot", "tile", "bar"), measure, ...)

Arguments

object	A Crunch variable or cube aggregation
...	additional plotting arguments
plot_type	One of "dot", "tile", or "bar" which indicates the plot family you would like to use. Higher dimensional plots add color coding or facets depending on the dimensionality of the data.
measure	The measure you wish to plot. This will usually be "count", the default but can also be ".unweighted_counts" or any other measure stored in the cube. If omitted, autoplot will select the first measure appearing in the data.

Value

A ggplot object.

---

Collect a Crunch dataset from the server

Description

This function brings a Crunch dataset into memory so that you can work with the data using R functions. Since this can create a long running query it is recommended that you try to filter the dataset down as much as possible before running collect().

Usage

## S3 method for class 'CrunchDataset'
collect(x, ...)

## S3 method for class 'GroupedCrunchDataset'
collect(x, ...)

Arguments

x	A Crunch Dataset
...	Other arguments passed to crunch::as.data.frame()

Details

When collecting a grouped CrunchDataset, the grouping will be preserved.

Value

A tbl_df or grouped_df

Examples

## Not run: 
ds %>%
   group_by(cyl) %>%
   select(cyl, gear) %>%
   collect()

## End(Not run)

---

Filter a Crunch dataset

Description

This function applies a CrunchLogicalExpression filter to a CrunchDataset. It's a "tidy" way of doing ds[ds$var == val,].

Usage

## S3 method for class 'CrunchDataset'
filter(.data, ..., .preserve = FALSE)

Arguments

.data	A CrunchDataset
...	filter expressions
.preserve	Relevant when the .data input is grouped. If .presrve = FALSE (the default), the grouping structure is recalculated based on the resulting data, otherwise the grouping is kept as is.

Value

.data with the filter expressions applied.

Examples

## Not run: 
ds %>%
   select(cyl, gear) %>%
   filter(cyl > 4) %>%
   collect()

## End(Not run)

---

Filter a Crunch dataset (deprecated)

Description

This function is deprecated, use filter() instead. Applies a CrunchLogicalExpression filter to a CrunchDataset. It's a "tidy" way of doing ds[ds$var == val,].

Usage

## S3 method for class 'CrunchDataset'
filter_(.data, ..., .dots)

Arguments

.data	A CrunchDataset
...	filter expressions
.dots	More dots!

Value

.data with the filter expressions applied.

---

Group-by for Crunch datasets

Description

group_by() sets grouping variables that affect what summarize() computes. ungroup() removes any grouping variables.

Usage

## S3 method for class 'CrunchDataset'
group_by(.data, ..., .add = FALSE)

## S3 method for class 'CrunchDataset'
ungroup(x, ...)

Arguments

.data	For group_by(), a Crunch Dataset
...	references to variables to group by, passed to dplyr::group_by_prepare()
.add	Logical: add the variables in ... to any existing grouping variables, or replace them (the default).
x	For ungroup(), a Crunch Dataset

Details

Note that group_by() only supports grouping on variables that exist in the dataset, not ones that are derived on the fly. dplyr::group_by() supports that by calling mutate() internally, but mutate is not yet supported in crplyr.

Value

group_by() returns a GroupedCrunchDataset object (a CrunchDataset with grouping annotations). ungroup() returns a CrunchDataset.

Examples

## Not run: 
ds %>%
   group_by(cyl) %>%
   select(cyl, gear) %>%
   collect()

## End(Not run)

---

A Crunch Dataset "Grouped By" Something

Description

This is a subclass of crunch::CrunchDataset that has a field for recording "group_by" expressions.

Examples

## Not run: 
ds <- loadDataset("Your dataset name")
class(ds) ## "CrunchDataset"
grouped_ds <- group_by(ds, var1)
class(grouped_ds) ## "GroupedCrunchDataset"

## End(Not run)

---

Mutate Crunch datasets (not implemented)

Description

Just a method that returns a nicer error message. mutate() hasn't been implemented yet. You can, however, derive expressions on the fly in summarize().

Usage

## S3 method for class 'CrunchDataset'
mutate(.data, ...)

Arguments

.data	A crunch Dataset
...	Other arguments, currently ignored

---

Select columns from a Crunch dataset

Description

This function uses "tidy select" methods of subsetting the columns of a dataset. It's another way of doing ds[,vars].

Usage

## S3 method for class 'CrunchDataset'
select(.data, ...)

Arguments

.data	A CrunchDataset
...	names of variables in .data or other valid selection functions, passed to tidyselect::vars_select()

Value

.data with only the selected variables.

Examples

## Not run: 
ds %>%
   select(contains("ear")) %>%
   filter(gear > 4) %>%
   collect()

## End(Not run)

---

Aggregate a Crunch dataset

Description

This is an alternate interface to crunch::crtabs() that, in addition to being "tidy", makes it easier to query multiple measures at the same time.

Usage

## S3 method for class 'CrunchDataset'
summarise(.data, ...)

Arguments

.data	A CrunchDataset
...	named aggregations to include in the resulting table.

Details

Note that while mutate() is not generally supported in crplyr, you can derive expressions on the fly in summarize().

Value

A tbl_crunch_cube or cr_tibble of results. This subclass of tibble allows ggplot2::autoplot to work, but can get in the way in some tidyverse operations. You may wish to convert to a tibble using as_tibble().

Examples

## Not run: 
ds %>%
    filter(cyl == 6) %>%
    group_by(vs) %>%
    summarize(hp=mean(hp), sd_hp=sd(hp), count=n())

## End(Not run)

---

Crunch ggplot theme

Description

Style ggplots according to Crunch style.

Usage

theme_crunch(base_size = 12, base_family = "sans")

Arguments

base_size	Base text size
base_family	Base text family

---

Return the unweighted counts from summarize

Description

This function allows you to return the unweighted counts from a Crunch dataset or grouped crunch dataset. It can only be used from within a summarise() call. If your dataset is unweighted, then unweighted_n() is equivalent to n().

Usage

unweighted_n()

Examples

## Not run: 
ds %>%
   group_by(cyl) %>%
   summarize(
       raw_counts = unweighted_n(),
       mean = mean(wt)
   )

## End(Not run)

---