Read and Manipulate Data from 'RiverWare'

A tool to read and manipulate data generated from 'RiverWare'(TM) <> simulations. 'RiverWare' and 'RiverSMART' generate data in "rdf", "csv", and "nc" format. This package provides an interface to read, aggregate, and summarize data from one or more simulations in a 'dplyr' pipeline.


Release (v0.6.1) Development
Travis-CI Build Status Travis-CI Build Status
AppVeyor Build Status AppVeyor Build Status
codecov codecov

RWDataPlyr is a tool to read and manipulate data generated from RiverWareTM simulations in rdf, csv, and nc formats and work with those data in a dplyr pipeline. It provides functions to gather, aggregate, and summarize data from multiple RiverWare simulations, i.e., scenarios.


RWDataPlyr can be installed from GitHub, and we suggest building the vignette.

# install.packages("devtools")
devtools::install_github('BoulderCodeHub/RWDataPlyr', build_vignettes = TRUE)


RWDataPlyr provides at least three workflows for reading and using RiverWare data:

  1. Reading and manipulating a single scenario
    • Fast
    • Best for inspecting a single slot
    • If comparing scenarios, must manually repeat for each scenario
    • Relies on read_rdf() and read_rw_csv()
  2. Summarizing multiple slots of data from a single scenario
    • Repeatable; allows user to process many slots at once
    • Best for producing "polished" analyses of a single scenario
    • Relies on rdf_aggregate() and user specified rwd_agg object
  3. Aggregating and summarizing many scenarios
    • Repeatable; allows user to process many slots for many scenarios at once
    • Repeats summary of a single scenario on multiple scenarios and combines results together
    • Relies on rw_scen_aggregate() and user specified rwd_agg object

Check out the workflow vignette for more details:

vignette("rwdataplyr-workflow", package = "RWDataPlyr")


  • 2018-06-07: version 0.6.1 available
  • 2018-04-10: version 0.6.0 available
  • 2017-05-26: version 0.5.0 available
  • 2016-11-01: version available. The package is now actually called RWDataPlyr.
  • 2016-10-20: version 0.4.1 available
  • Previous versions were originally available as the RWDataPlot package
    • 2016-07-13: version 0.4 available
    • 2016-03-22: version 0.3 available
    • 2015-07-01: version 0.2 available
    • 2014-09-16: working to create an R Package from existing code.


This software is in the public domain because it contains materials that originally came from the U.S. Bureau of Reclamation, an agency of the United States Department of Interior.

Although this code has been used by Reclamation, no warranty, expressed or implied, is made by Reclamation or the U.S. Government as to the accuracy and functioning of the program and related program material nor shall the fact of distribution constitute any such warranty, and no responsibility is assumed by Reclamation in connection therewith.

This software is provided "AS IS."


RWDataPlyr 0.6.2

Released August 15, 2018

Minor Changes

  • The oFile argument in getDataForAllScens() now defaults to NULL, so that this function conforms to CRAN policies regarding not writing to the user's home file space by default. This should not cause any backwards compatability issues since all older code will explicitly specify the oFile argument.
  • The default file argument in rwd_agg_template() is now empty, so the user must specify the file explicitly for it to be created, also to conform to the same CRAN policy.
  • Tests, examples, and the vignetter were updated to respect this policy, by only writing to the tempdir(), when necessary.

RWDataPlyr 0.6.1

Released June 8, 2018

Minor new features

  • A new function, rdf_to_rwtbl2(), was added to try and improve performance of rdf_aggregate(), and rw_scen_aggregate(). (#85) These functions were shown to be about 6x - 7x slower than getDataForAllScens() for the same aggregation. The new function calls a C++ version that creates the tbl_df, while maintaining the same information as rdf_to_rwtbl(). The API for the new function is slighly different; the first argument is now a path to an rdf file, rather than an already read in rdf object. Otherwise, the same options are available. The C++ version of rdf_to_twbl() is about 20x faster than the R version. However, rdf_aggregate(), and rw_scen_aggregate() are still about 4x slower than getDataForAllScens(), indicating that there is still some necesary work to get closer speeds between the two functions (#90). As part of this work, the following modifications to other functions were made:
    • rw_scen_aggregate() and rdf_aggregate() gain cpp arguments. By default, rdf_to_rwtbl2() is used, but rdf_to_rwtbl() can be foreced by setting cpp = FALSE.
    • read_rdf() and read.rdf() gained an rdf argument. If TRUE (default), it returns an rdf object, otherwise it returns a character vector.
    • Deprecate rdf_to_rwtbl() in favor of rdf_to_rwtbl2()
    • In rdf_to_rwtbl(), scenario is coerced into a character. Typically this is a character, but it was previously left as numeric if specified as a numeric. For easier compatability with C++ and comparsion between rdf_to_rwtbl() and rdf_to_rwtbl2(), it's now always a character.
  • rdf_aggregate() and rw_scen_aggregate() gain a verbose parameter to print out the status of processing multiple scenarios, rdfs, and slots. (#82)
  • A new function (rwd_agg_template()) to create a blank template (or with examples: examples = TRUE) csv file to use to create rwd_agg objects. (#78)
  • A new function (read_rwd_agg()) to read in csv files as rwd_agg objects. (#70)

Bug fixes

  • Improved read_rdf() error messages (#86)
  • rw_scen_aggregate() will now work with unnamed scenarios and NULL scen_names arguments. (#81)
  • rwslot_* functions now error if the data passed to them are not regular (January - December or October - September) (#83)
    • As part of this, the matrix returned by rdf_get_slot() now has a "timespan" attribute that corresponds to the start and end values of the rdf.
  • Updated documentation for columns returned by rdf_aggregate() and rw_scen_aggregate(). Now it is clear that the Timestep columns is not returned. (#84)

RWDataPlyr 0.6.0

Released April 10, 2018

RWDataPlyr v0.6.0 includes a major revamping of how scenarios are processed and defines several new classes.

Major new features

  • New functions rw_scen_aggregate() and rdf_aggregate() added to upgrade existing getDataForAllScens() function, which processes multiple scenarios at a time. (#51)

    • These two functions rely on the new rwd_agg class, which upgrades the exising "slot aggregation list". The advantage of the new class is a much more flexible way to summarize and aggregate RiverWare slots. (#68)
      • This class includes methods for rbind(), cbind(), as, and is.
    • Helper functions to print all slots that exist in the tibble (rwtbl_slot_names()), get the original scenario folders (rwtbl_get_scen_folder()), and get RiverWare slots names from the saved variable name (rwtbl_var_to_slot()) were added. (#50)
  • getDataForAllScens() now always returns data invisibly, so the retFile arguement is deprecated. This function is also deprecated in favor of rw_scen_aggregate() and/or rdf_aggregate(). (#66)

  • Formalized the list returned by createSlotAggList() as a slot_agg_list class. Created applicable constructor, which deprecates createSlotAggList(). Includes print(), summary() and is./is_ methods and functions. (#67)

    • Added check in creation to ensure all variables are unique (if specified) (#64, #62)
    • However, slot_agg_list objects work with the deprecated getDataForAllScens(), so rwd_agg objects are preferable.
  • Addedread_rw_csv() to read csv files created from RiverWare or RiverSMART. (#30)

  • Addedrdf_to_rwtbl() to convert rdf objects (read in from read.rdf()) to tibble objects. (#30)

  • Added rdf class

    • read.rdf() now returns an object with an rdf class
    • Functions that expected rdf lists now expect an rdf object, e.g., rdfSlotToMatrix().
  • Switched to snake_case for all new functions and replaced existing functions with snake_case versions. (#76)

    Old Function New Function
    getSlotsInRdf() rdf_slot_names()
    makeAllScenNames() rw_scen_gen_names()
    rdfSlotToMatrix() rdf_get_slot()
    sumMonth2Annual() rwslot_annual_sum()
    getTimeSpan() rdf_get_timespan()
    getWYFromYearmon() ym_get_wateryear()
    getMaxAnnValue() rwslot_annual_max()
    getMinAnnValue() rwslot_annual_min()
    flowWeightedAvgAnnConc() rwslot_fwaac()
    • For flowWeightedAvgAnnConc(), because it is rarely used it was converted to an internal function: trace_fwaac(). Then created rwslot_fwaac() that is exported and follows the input/output format of rwslot_annual_min() and rwslot_annual_max().
  • Added the rwdataplyr-workflow vignette. Browse with vignette("rwdataplyr-workflow", package = "RWDataPlyr").

Minor improvements

  • For monthly data, getDataForAllScens() will use full month name in the Month column. This change could break existing code if there are checks for particular months. (#20)
  • Thresholds < 1 will now work for aggregation methods that compute <=/>= for slot_agg_list and getDataForAllScens(). (#14)
  • Removed dependency on reshape2 and converted to use dplyr where appropriate.
  • Got rid of read.rdf()'s original implementation. Now read.rdf2() (the faster implementation) is named read.rdf() and read.rdf2() is deprecated. (#63)
    • read.rdf() now works with rdf files that contain scalar slots (#52)
    • read_rdf() is added as an alias.

RWDataPlyr 0.5.0

Released May 26, 2017

Major new features

  • Modified all aggregation methods that check if the data are <, <=, >, or >= a threshold. Previously, these aggregation methods returned either 0 (FALSE) or 100 (TRUE), now they return 1 for TRUE. This allows us to keep data in its original state before plotting with scales::percent, and keep us from having to multiply/divide by 100 if we compute other averages outside of RWDataPlyr. However, this will cause any plotting code that is expecting percentages between 0 and 100 instead of between 0 and 1 to break (or at least look weird). (#53)
  • Added findAllSlots boolean parameter to getDataForAllScens()
    • If this is TRUE an error will post if one or more of the slots cannot be found. If it is FALSE, then it will fill in the data frame with -99, but not fail. (#38)
  • New function: makeAllScenNames() will create scenario names for vectors of multiple dimensions.
  • New function: getWYFromYearmon() will take a yearmon object and determine the water year it falls into.
  • Added WYMaxLTE as a valid aggregation method. This will compare the maximum value in a water year to a threshold and determine if it is less than or equal to the threshold.

Minor improvements

  • Better error messages in createSlotAggList() (#45)
  • Improved the error message if an invalid aggregation method is passed to processSlots() (#42)
  • Ensure that getDataForAllScens() (and interal function processSlots()) work with rdf files that only include one trace of data (#40)
  • getDataForAllScens() through processSlots() now returns the same type/class for Year and Variable columns for both annual and monthly data. Year is a numeric and Variable is always a character. This could affect code that does not read the data frame back in and assumes that Variable is a factor. (#54)

Under-the-hood improvements

  • Improved tests for createSlotAggList()
  • Added tests for RWDataPlyr:::processSlots()
  • Now check for invalid aggregation methods in createSlotAggList(), so it will not wait until processSlots() is called to throw the error.

Reference manual

It appears you don't have a PDF plugin for this browser. You can click here to download the reference manual.


0.6.4 by Alan Butler, a year ago

Report a bug at

Browse source code at

Authors: Alan Butler [aut, cre] , Cameron Bracken [aut]

Documentation:   PDF Manual  

CC0 license

Imports data.table, dplyr, methods, stats, tibble, tidyr, tools, utils, feather, xts, zoo, Rcpp

Suggests bookdown, knitr, rmarkdown, testthat, covr

Linking to Rcpp

See at CRAN