Data frames with indexes

Description

data frames for which observations are defined by two (potentialy nested) indexes and for which series have thefore a natural tabular representation

Usage

dfidx(
  data,
  idx = NULL,
  drop.index = TRUE,
  as.factor = NULL,
  pkg = NULL,
  fancy.row.names = FALSE,
  subset = NULL,
  idnames = NULL,
  shape = c("long", "wide"),
  choice = NULL,
  varying = NULL,
  sep = ".",
  opposite = NULL,
  levels = NULL,
  ranked = FALSE,
  name,
  position,
  sort = TRUE,
  drop.unused.levels = TRUE,
  ...
)

Arguments

Details

Indexes are stored as a data frame column in the resulting dfidx object

Value

an object of class "dfidx"

Author(s)

Yves Croissant

Examples

# the first two columns contain the indexes
mn <- dfidx(munnell)

# explicitely indicate the two indexes using either a vector or a
# list of two characters
mn <- dfidx(munnell, idx = c("state", "year"))
mn <- dfidx(munnell, idx = list("state", "year"))

# rename one or both indexes
mn <- dfidx(munnell, idnames = c(NA, "period"))

# for balanced data (with observations ordered by the first, then
# by the second index

# use the name of the first index
mn <- dfidx(munnell, idx = "state", idnames = c("state", "year"))

# or an integer equal to the cardinal of the first index
mn <- dfidx(munnell, idx = 48L, idnames = c("state", "year"))

# Indicate the values of the second index using the levels argument
mn <- dfidx(munnell, idx = 48L, idnames = c("state", "year"),
            levels = 1970:1986)

# Nesting structure for one of the index
mn <- dfidx(munnell, idx = c(region = "state", president = "year"))

# Data in wide format
mn <- dfidx(munnell_wide, idx = c(region = "state"),
            varying = 3:36, sep = "_", idnames = c(NA, "year"))

# Customize the name and the position of the `idx` column
dfidx(munnell, position = 3, name = "index")

Index for dfidx

Description

The index of a dfidx is a data frame containing the different series which define the two indexes (with possibly a nesting structure). It is stored as a "sticky" data frame column of the dfidx object and is also inherited by series (of class 'xseries') which are extracted from a dfidx object.

Usage

idx(x, n = NULL, m = NULL)

## S3 method for class 'dfidx'
idx(x, n = NULL, m = NULL)

## S3 method for class 'idx'
idx(x, n = NULL, m = NULL)

## S3 method for class 'xseries'
idx(x, n = NULL, m = NULL)

## S3 method for class 'idx'
format(x, size = 4, ...)

Arguments

Details

idx is defined as a generic with a dfidx and a xseries method.

Value

a data frame containing the indexes or a series if a specific index is selected

Author(s)

Yves Croissant

Examples

mn <- dfidx(munnell, idx = c(region = "state", president = "year"))
idx(mn)
gsp <- mn$gsp
idx(gsp)
# get the first index
idx(mn, 1)
# get the nesting variable of the first index
idx(mn, 1, 2)

Get the name and the position of the index column

Description

This function extract the names of the indexes (along with the position of the idx column) or the name of a specific index

Usage

idx_name(x, n = 1, m = NULL)

## S3 method for class 'dfidx'
idx_name(x, n = NULL, m = NULL)

## S3 method for class 'idx'
idx_name(x, n = NULL, m = NULL)

## S3 method for class 'xseries'
idx_name(x, n = NULL, m = NULL)

Arguments

Value

if n is NULL, a named integer which gives the position and the name of the idx column in the dfidx object, otherwise, a character of length 1

Author(s)

Yves Croissant

Examples

mn <- dfidx(munnell, idx = c(region = "state", president = "year"))
# get the position of the idx column
idx_name(mn)
# get the name of the first index
idx_name(mn, 1)
# get the name of the second index
idx_name(mn, 2)
# get the name of the nesting variable for the second index
idx_name(mn, 2, 2)

Methods for dfidx

Description

A dfidx object is a data frame with a "sticky" data frame column which contains the indexes. Specific methods of functions that extract lines and/or columns of a data frame are provided : [, [[, $,⁠[<-⁠, ⁠[[<-⁠ and ⁠$<-⁠. Moreover, methods are provided for base::transform and base::subset in order to easily generate new variables and select some rows and columns of a dfidx oject. An organize function is also provided to sort a dfidx object using one or several series.

Usage

## S3 method for class 'dfidx'
x[i, j, drop]

## S3 method for class 'dfidx'
as.data.frame(x, row.names = NULL, optional = FALSE, ...)

## S3 method for class 'dfidx'
print(x, ..., n = NULL)

## S3 method for class 'dfidx'
head(x, n = NULL, ...)

## S3 method for class 'dfidx'
x[[y]]

## S3 method for class 'dfidx'
x$y

## S3 replacement method for class 'dfidx'
object$y <- value

## S3 replacement method for class 'dfidx'
object[[y]] <- value

## S3 method for class 'xseries'
print(x, ..., n = NULL)

## S3 method for class 'idx'
print(x, ..., n = NULL)

## S3 method for class 'dfidx'
mean(x, ...)

## S3 method for class 'dfidx'
transform(`_data`, ...)

## S3 method for class 'dfidx'
subset(x, subset, select, drop = FALSE, drop.unused.levels = TRUE, ...)

organize(x, ...)

Arguments

Value

as.data.frame and mean return a data.frame, [[ and $ a vector, [ either a dfidx or a vector, ⁠$<-⁠ and ⁠[[<-⁠ modify the values of an existing column or create a new column of a dfidx object. transform, subset and organize return a dfidx object. print is called for its side effect.

Author(s)

Yves Croissant

Examples

mn <- dfidx(munnell)
# extract a series (returns as a xseries object)
mn$gsp
# or
mn[["gsp"]]
# extract a subset of series (returned as a dfidx object)
mn[c("gsp", "unemp")]
# extract a subset of rows and columns
mn[mn$unemp > 10, c("utilities", "water")]
# dfidx, idx and xseries have print methods as (like tibbles), a n
# argument
print(mn, n = 3)
print(idx(mn), n = 3)
print(mn$gsp, n = 3)
# a dfidx object can be coerced to a data.frame
as.data.frame(mn)
# transform, subset and organize are usefull methods/function to
# create new series, select a subset of lines and/or columns and to
# sort the `dfidx` object using one or several series
transform(mn, gsp70 = ifelse(year == 1970, gsp, 0))
subset(mn, gsp > 200000, select = c("gsp", "unemp"))
subset(mn, 1:20, select = c("gsp", "unemp"))
organize(mn, year, unemp)

model.frame and model.matrix methods for dfidx objects

Description

Specific model.frame and model.matrix are provided for dfidx objects. This leads to an unusual order of arguments compared to the usage. Actually, the first two arguments of the model.frame method are a dfidx and a formula and the only main argument of the model.matrix method is a dfidx which should be the result of a call to the model.frame method, i.e. it should have a terms attribute.

Usage

## S3 method for class 'dfidx'
model.frame(
  formula,
  data = NULL,
  ...,
  lhs = NULL,
  rhs = NULL,
  dot = "previous",
  alt.subset = NULL,
  reflevel = NULL,
  balanced = FALSE
)

## S3 method for class 'dfidx'
model.matrix(object, ..., lhs = NULL, rhs = 1, dot = "separate")

## S3 method for class 'dfidx_matrix'
print(x, ..., n = NULL)

Arguments

Value

a dfidx object for the model.frame method and a matrix for the model.matrix method.

Author(s)

Yves Croissant

Examples

mn <- dfidx(munnell)
mf <- model.frame(mn, gsp ~ privatecap | publiccap + utilities | unemp + labor)
model.matrix(mf, rhs = 1)
model.matrix(mf, rhs = 2)
model.matrix(mf, rhs = 1:3)

Productivity in the United States

Description

a panel data of 48 American States for 17 years, from 1970 to 1986

Usage

munnell

munnell_wide

Format

a data frame containing:

• state: the state

• year: the year

• region: one of the 9 regions of the United States

• president: the name of the president for the given year

• publiccap: public capital stock

• highway: highway and streets

• water: water and sewer facilities

• utilities: othe public building and structures

• privatecap: private capital stock

• gsp: gross state product

• labor: labor input measured by the employment in non–agricultural payrolls

• unemp: state unemployment rate

An object of class tbl_df (inherits from tbl, data.frame) with 48 rows and 36 columns.

Source

Online complements to Baltagi (2001): https://www.wiley.com/legacy/wileychi/baltagi/ Online complements to Baltagi (2013): https://bcs.wiley.com/he-bcs/Books?action=resource&bcsId=4338&itemId=1118672321&resourceId=13452

References

Baltagi BH (2001). Econometric Analysis of Panel Data, 3rd edition. John Wiley and Sons ltd.

Baltagi BH (2013). Econometric Analysis of Panel Data, 5th edition. John Wiley and Sons ltd.

Baltagi BH, Pinnoi N (1995). “Public capital stock and state productivity growth: further evidence from an error components model.” Empirical Economics, 20, 351-359.

Munnell A (1990). “Why Has Productivity Growth Declined? Productivity and Public Investment.” New England Economic Review, 3–22.

Fold and Unfold a dfidx object

Description

fold_idx takes a dfidx object, includes the indexes as stand alone columns, remove the idx column and return a data frame, with an ids attribute that contains the informations about the indexes. fold_idx performs the opposite operation.

Usage

unfold_idx(x)

fold_idx(x, pkg = NULL, sort = FALSE)

Arguments

Value

a data frame for the unfold_dfidx function, a dfidx object for the fold_dfidx function

Author(s)

Yves Croissant

Examples

mn <- dfidx(munnell, idx = c(region = "state", "year"), position = 3, name = "index")
mn2 <- unfold_idx(mn)
attr(mn, "ids")
mn3 <- fold_idx(mn2)
identical(mn, mn3)