Functions and S3 classes for time indexes and time indexed series, which are compatible with FAME frequencies.
Changes in Version 1.37
o Adjust some more examples o Fix couldBeTi
Changes in Version 1.36
o Adjust some examples
Changes in Version 1.35
o Remove makevars files to comply with cran o add in check in baseYmd to make sure tif is a whole number
Changes in Version 1.34
o Add tis_init.c to eliminate note in cran checks
Changes in Version 1.33
o adjusted convert to properly account for period bounderies when using businessday
data types from fame
o Removed format.POSIXct, format.POSIXlt, format.default from being exported
Changes in Version 1.32
o tisPlot() and scatterPlot() parameters xExpandBy, leftExpandBy, rightExpandBy and yExpandBy can now be a vector of two numbers to expand the data range asymmetrically.
o tisFromCsv() was bombing when dateFormat = "excel" was given.
o created a non-exported is.numeric.ti() that returns FALSE and explicitly exported as.Date.jul and as.Date.ti to prevent the zoo package's reimplementation of the generic as.Date from incorrectly sending ti arguments to zoo::as.Date.numeric().
Changes in Version 1.31
o Added 'offset' arg to as.Date.ti, and 'origin' arg to as.Date.jul. Request of Andre Mikulec.
o Fixed minor bug in barplot.tis()
Changes in Version 1.30
o Added tis methods for head() and tail().
Changes in Version 1.29
o Removed aggregate.ts() function, which was a modified version of the same-named function in package:stats. Also modified aggregate.tis() to prefer a local version of aggregate.ts to the one in package:stats.
o Registered S3 method t.tis()
Changes in Version 1.28
o Added tz argument to tisFromCsv() and added logic to handle date strings that can't be turned into valid dates. Thanks again to Ying Leung for reporting the problem that led to the discovery of a bug.
o Added naLoses argument to mergeSeries(), with default value of FALSE for backward compatibility.
Changes in Version 1.27
o Fixed bug in tisFromCsv() reported by Ying Leung. Also added chopNAs parameter to that function.
o Changed name of ymd.xy() to ymdXy() to keep the CRAN checker happy.
Changes in Version 1.26
o now exports the barplot2() function called by barplot.tis
Changes in Version 1.25
o new capitalize() function moved here from the another package.
o new functions ymd.xy() and ymdShade() generalize the old nber.xy() (now defunct) and nberShade() functions to shade arbitrary date ranges on tisPlots, not just NBER recessions.
o changed names of "nber" arguments to tisPlot() to reflect the more general date range shading functionality.
Changes in Version 1.24
o freq2tif() gave wrong answer for frequency = 261, business day freq.
o tierChart(vlines.month) now draws vertical line after the last day of the month, rather than after the first day of the month.
Changes in Version 1.23
o added some support for IDate objects from the data.table package.
o added isBusinessDay() function at the suggestion of Adam Hogan.
o fixed a bug in inferDate() that was causing tisFromCsv() to misbehave
Changes in Version 1.22
o defined some variables locally in nberShade.ggplot() to avoid spurious R CMD check warnings.
Changes in Version 1.21
o changed name of as.list.keepClass() to asClassyList() to placate CRAN
Changes in Version 1.20
o fixed bug in assignList()
Changes in Version 1.19
o added xOffset argument to tisPlot()
o fixed a bug in nberShade.ggplot() found by Matthew McCormick. Thanks, Matt.
Changes in Version 1.18
o added a tis method for the generic barplot() function
o moved tierChart functions from local frb package to tis package
Changes in Version 1.17
o ymdToTi() (an internal function in ti.R) for frequencies bmnovember, qoctober and qnovember was giving results that were off by one period, causing convert() to give wrong results when aggregating to those frequencies. Thanks again to Abiel Reinhart for reporting the problem.
o changed "env = envir" to "envir = envir" in several places to avoid package check complaints about partial matches
Changes in Version 1.16
o more bugs in convert() reported by Abiel Reinhart
Changes in Version 1.15
o fixed bug in convert() reported by Abiel Reinhart
Changes in Version 1.14
o added as.POSIXct.jul and as.POSIXct.ti methods
Changes in Version 1.13
o make "[<-.tis" a bit smarter when only a single index is supplied
Changes in Version 1.12
o fixed bugs in "[.tis" and "[<-.tis" reported by Abiel Reinhart.
Changes in Version 1.11
o added 'zero' argument to naWindow
Changes in Version 1.10
o added methods for generic function xtfrm for classes jul, ssDate, ti and tis.
o print.ti() and print.tis() can now suppress the 'class: ti' and 'class: tis' strings, respectively, at the end.
Changes in Version 1.9
o convert() uses tapply() which calls as.factor() on a ti object. For some reason, this results in NA's in R 2.10.0. Unclassing the ti before using it as a parameter to tapply fixed it. Why factor() gives NA's for ti objects is a subject for further investigation.
Changes in Version 1.8
o Allow fortify.tis to handle proper naming when a univariate array is converted into a tis object. Contributed by Trevor Davis.
Changes in Version 1.7
o Tweaked nberShade.ggplot to work with a greater range of ggplot objects, and improved examples for nberShade and fortify.tis documentation. Contributed by Trevor Davis.
Changes in Version 1.6 (contributed by Trevor Davis)
o made nberShade() generic and added a method for ggplot objects
o added fortify.tis(), a tis-specific method for the fortify() generic in ggplot2
Changes in Version 1.5
o added options to how nberDates() and nber.xy() handle an ongoing recession
Changes in Version 1.4
o fixed nberDates() to show latest recession
Changes in Version 1.3
o cleaned up some documentation to incorrect package references and references to Fed-specific stuff.
o added "tif" parameter to today() with default value "daily"
Changes in Version 1.2
o better handling of ti objects with some NA values. Thanks to Chuck Mikolajczak for pointing out the problem.
o the observed attribute of a tis can now be one of "beginning", "ending", "summed", "averaged", "annualized", "high", or "low". The last two are new, and denote series for which only the high or low value over a period is recorded. The convert() function can work with the new attributes.