gWidgets (version 0.0-54.2)

gtable: Constructor for widget to display tabular data

Description

This widget displays either a vector, matrix or data frame in a tabular format. The main usage is for user selection of a row or rows.

Usage

gtable(items, multiple = FALSE, chosencol = 1, icon.FUN = NULL,
filter.column = NULL, filter.labels = NULL, filter.FUN = NULL, 
handler = NULL, action = NULL, container = NULL, ..., toolkit = guiToolkit())

Arguments

items

A vector, matrix or data frame to be displayed. A vector and matrix is coerced into a data frame.

multiple

A logical. If TRUE multiple lines can be selected

chosencol

By default, only the value in this column is returned by the svalue method.

icon.FUN

If given, this function is applied to the data frame to be shown. It should return a vector of stock icon names

filter.column

If not NULL a filter by droplist is given which can be used to filter the displayed values shown using the values in this column.

filter.labels

If more complex filtering is desired then this argument populates the values of the combobox and the filter.FUN argument is used to specify a function to interpret these values.

filter.FUN

Either a function with signature (obj, filter.by) to specify a vector of logical values indicating which rows should be shown or the character "manual" in which filtering is done directly through the visible method and not through a popup box

handler

Called on a double click event

action

Passed to handler

container

Optional container to attach widget to

…

ignored

toolkit

Which GUI toolkit to use

Details

The svalue method returns the selected value(s). By default, only the value(s) in the chosen column are returned. Use the argument drop=FALSE to return the entire row. To return the row index, use the argument index=TRUE. This index refers to the whole data store, not just the visible portion when filtering is being used.

The "[" notion treats the object like a data frame. When filtering, this notation refers to the entire data frame, not the visible data frame. The comment about the returned index by svalue can be described by the fact that obj[svalue(obj,index=TRUE), ] should be the same value as svalue(obj).

Assignment via "[<-" is possible with limitations imposed by the toolkits. The graphical display of tabular matter is usually done, similar to data frames, in terms of columns each having the same type. For some toolkits, all values are converted to characters, for others, the type must be maintained. In R, coercion of types may occur when assigning to a data frame, but this won't be so with the underlying toolkit widget. To be portable across toolkits, the column type should not change during assignment, nor should the number of rows be reduced.

In particular, assignment with "[<-" for factors can cause warnings if the values are not in the factor's levels. When the value being assigned is a matrix there is a coercion to a data frame which may change the type.

The visible and visible<- methods refer to which rows of the data store are visible in the widget. These are specified by a vector of class logical. This may be used when there is filtering, not sorting. The value returned by svalue is a logical vector of length given by the number of rows of the data store, with TRUE indicating that the row is displayed. When setting the visibility of a row through svalue the vector of values should have the same length as the number of rows, otherwise recycling occurs. (So visible(obj) <- TRUE will work to display all the rows.)

The length method returns the length of the underlying data store.

The dim method returns the dimension of the underlying data store.

The names method returns the names of the underlying data store.

The names<- method can be used to set the names of the underlying data store and the values displayed in the column headers of the widget.

Row names are ignored in the display of this widget.

A single click is used for selection of a value.

The addHandlerDoubleclick handler can be used to define a callback to respond to a double click event.

See Also

See also gtree for displaying tree-like data and gdf for tabular data meant to be edited

Examples

# NOT RUN {
  ## example to select CRAN mirror
  m <- getCRANmirrors()[,c(1,4)]
  setCRAN <- function(URL) { ## see chooseCRANmirror
    repos = getOption("repos")
    repos["CRAN"] <- gsub("/$", "", URL)
    options(repos=repos)
  }
  w <- gwindow("gtable example",width=400)
  gp <- ggroup(horizontal=FALSE, container=w)
  tab <- gtable(m, chosencol = 2, container=gp, expand=TRUE,
       handler = function(h,...) setCRAN(svalue(h$obj)))
  bg <- ggroup(container=gp)
  addSpring(bg)
  gbutton("dismiss", container=bg, handler = function(h,...) dispose(w))


  ## an example with icons.
  ## Select variables from a data frame


  ## find icons by class
  icon.FUN <- function(items) {
    dfName <- svalue(cb)
    df <- try(get(dfName, envir=.GlobalEnv), silent=TRUE)
    if(inherits(df,"try-error"))
      return(rep(NULL,dim(items)[1]))
    if(is.data.frame(items))
      items <- items[,1, drop=TRUE]
    sapply(items, function(i) {
      class(df[,i])[1]
    })
  }
  ## list data frames in an environment
  lsDF <- function(envir=.GlobalEnv) {
    varNames <- ls(envir=envir)
    dfs <- sapply(varNames, function(i) inherits(get(i,envir=envir),"data.frame"))
    varNames[dfs]
  }
  ## set up GUI
  w <- gwindow("Select variables",width=250)
  g <- ggroup(horizontal=FALSE, container=w)
  l <- glabel("Data frame", container=g)
  cb <- gcombobox(lsDF(), container=g)
  blankDF = data.frame(variables=character(0), stringsAsFactors=FALSE)
  tbl <- gtable(blankDF, icon.FUN=icon.FUN, container=g, expand=TRUE)
  ## add handlers
  addHandlerChanged(cb, handler <- function(h,...) {
    dfName <- svalue(h$obj)
    dfNames <- names(get(dfName,envir=.GlobalEnv))
    tbl[,] <- data.frame(variables=dfNames, stringsAsFactors=FALSE)
  })
  addHandlerClicked(tbl, handler = function(h,...) {
    cat("Do something with",svalue(cb),"::", svalue(h$obj),"\n")
  })


# }