htsr-package

Pierre Chevallier

HydroSciences Laboratory (IRD, CNRS, Univ. Montpellier, IMT Mines Ales), France

v 2.1.4

Content

Origin
- Brief history
- Hydraccess
- Sqlite data base
Principles
Dependencies
- System library
- User library
- Special case of RODBC
File formats
- Sqlite data base
- Time-series
- Calibrations and discharge measurements
- Gaps
Functions
- Short cut fonction
- Data base functions
- File functions
- Hydro-meteo time-series
- Plot functions
- Micellaneous functions
Development main references
- Sqlite management
- R coding
Appendix
- Correspondence of the data base tables between Hydraccess and Sqlite htsr
- Fields of the main tables

Origin

Brief history

IRD and previously ORSTOM have developed quite early (in the 70ties) data bases for storing hydrological and meteorological time-series. During the 80ties the managing tools use home-made codes (Hydrom and Pluviom), which were in the late 90ties adapted with commercial applications like Microsoft Access or Oracle and more recently for web interface.

With the development of open access platforms and applications, in the one hand, and the need of more flexibility, in the other hand, it appears useful to build a set of functions able to manage hydro-meteo time-series independently of the operating system and of proprietary tools.

The htsr package was developed on the basis of the public domain data base management system Sqlite and on the hydro-meteo time-series data base application Hydraccess.

It works on Linux, Windows or MacOS platforms, with the free programming language R and the interface RStudio preliminary installed.

Hydraccess

Hydraccess was developed at IRD (https://www.ird.fr) by Philippe Vauchel. It is based on several tools provided by Microsoft Office (R) including Microsoft Access. It can be obtained from https://hybam.obs-mip.fr/fr/hydraccess-2/

Hydracces is available in French, English, Spanish, Portuguese and Russian.

Sqlite data base

Sqlite is an embedded data base on the public domain (https://en.wikipedia.org/wiki/SQLite). It is self-contained, and serverless, without preliminary configuration. It can be installed using the instructions given by https://sqlite.org/download.html. It can be used directly through many dedicated applications available for a large set of operating systems.

Principle

The Sqlite data base is organized with the same tables and fields as the Hydraccess structure, except that the table names are changed for practical reasons. A correspondence of these names in given in appendix.

In Hydraccess, two families or time-series coexist: “hydro data” (Type_Station = H) and “weather data” (Type_Station = M). The first one includes the main tables Cotes/WL, Debits/DI and Qualite/QU; the second one includes Pluies/PR and Meteo/WE. Other tables regarding discharge measurements or calibration are also affected to the “hydro data” family.

This distinction does not exist anymore in the Sqlite data base processes. That means that one should verify if no stations have the same name in the one and the other family. If it is the case the name must be changed before converting a Hydraccess data base in a Sqlite data base.

A time-series is always attached to a station, which describes its location and managing infos, and to a sensor, which is attached to the data type and to the used device. Both are indexed with a main id.

In Hydraccess:

The station id has no particular syntax; it associated to a number of description fields. Note that the station Name field is mandatory.
The capteur/sensor id has at most 5 digits and the first one defines the timestep of the variable: I for instantaneous, J for daily and M for monthly. By convention, a daily value is affected at 12:00:00 and a monthly value to the day 16 at 00:00:00. This convention remains for the Sqlite data base. The other 4 digits of the sensor id are free in Hydraccess. In the Sqlite data base the sensor id length is not any more limited, but it could be a good idea to conserve it. Other description fields are also associated with the sensors.

In brief, that means that a time-series record must have 4 objects:

a date-time;
a numeric value;
a station id factor;
a sensor id factor.

Practically it associates the two tables “station” and “sensor” with one of the five following: “water level”, “discharge”, “weather”, “precipitation” or “quality”.

The other associated tables can be used for intermediary calculations, like the computation of the discharge from the water level, or for more detailed information.

Dependencies

The following packages and their own dependencies are needed to apply the htsr package.

tidyverse, directlabels, editData, openair, Rcpp, RSQLite, RODBC, shiny, shinyFiles, terra, WriteXLS, fs

Special case of RODBC

Because Hydraccess is only configured for Microsoft Windows, the associated data bases work only on this platform. That means that the operation, which consists to convert a Hydraccess data base into a Sqlite data base must be done in this environment. The RODBC package is associated to the htsr package, but it can be only used in the Windows platform through the function d_convert_hydraccess. In particular, this function does not work on Linux.

Because a Hydraccess data base is proprietary depending, a preliminary configuration of the MS-Windows platform must be done:

Install “Microsoft Database Engine” (freely available on Microsoft web site).
In the administration tools, config the connection ODBC 32b with the Access data base to be used. An alternative consists to execute \windows\sysWOW64\odbcad32.exe.
Open the RStudio session in 32b (this can be precised in the RStudio general options ; RStudio must be re-launched for taking the change into account).

File formats

Sqlite data base

extension : .sqlite
SQLite data base with 38 tables.

Time-series

extension : .hts
Rdata file containing a “tibble” object, tstab, with 4 columns : Date, Value, Station, Sensor

> tstab
   Date                Value Station Sensor
   <dttm>              <dbl> <fct>   <fct> 
 1 2015-10-03 12:00:00  2.17 CKS2500 IQ

Gaps

extension : .gap
Rdata file containing a “tibble” object, ze, with 3 columns

> ze
# A tibble: 14 x 3
   date                valeur stacapt
   <dttm>               <dbl> <chr>  
 1 2015-10-03 17:00:00      1 _      
 2 2015-12-11 23:30:00      1 _

Functions

6 categories of functions are provided by the htsr package, with, in addition, one short-cut function. They are distinguished with a prefix character and listed below.

Infos on their uses are detailed in the on-line help or using in the console the command ?<FUNCTION_NAME> or help(<FUNCTION_NAME>).

Short-cut function

fc(file), shortcut for file.choose(file), which is very frequently needed.

Data base functions (prefix `d_` or `ds_`)

Data base tools

d_backup(fsq), backups a htsr sqlite data base.
d_compact(bd.sqlite), compacts a htsr sqlite data base
d_create(fsq, cr_table = TRUE, bku = TRUE), creates a htsr sqlite data base
ds_inventory_station() and ds_inventory_sensor(), produce the inventory of the stations, respectively sensors of a selected station, stored in a htsr sqlite data base.
ds_sensor(), creates, modifies or removes a sensor.
ds_station(), creates, modifies or removes a station.
d_table(fsq, table, op = "C", bku = TRUE) creates or removes a table

Data import/export

d_exp_discalib(fsq, sta, calib=TRUE, dism=TRUE), exports discharge measurements and calibrations from data base
ds_exp_hts(), extract hts files from a sqlite data base
d_imp_hts(fsq, file, table, bku = TRUE), imports a hts file into a data base
d_rem_hts(fsq, table, sta, sen, start = NA, end = NA), removes a htsr record from a data base

Conversion

d_convert_hydraccess(fsq, db.hydraccess), converts a full Hydraccess database into a new htsr sqlite database. NB: Only works in Windows environment with a 32b R session.
d_convert_weewx(fsq, db.weewx, sta, name_st, tzo = "CET", bku = TRUE), converts a weewx data base (http://weewx.com/) into a htsr sqlite base
d_convert_meteofrance_d(fmeteo), converts a Meteo France csv daily data file into a htsr sqlite base

File functions (prefix `f_`)

f_change_id(file, sta, sen, overwrite), changes station id or sensor id in a hts file.. Generates a file with the prefix nw_.
f_convert(file, form_start = "hts", form_end = "xlsx", ta = NA, sen = NA, output = NA, variable = NA), converts data-series file in another format: hts, xls/xlsx, csv) and viceversa.
f_csv_multivar(files, daily = TRUE, fileo = "fileo"), build a multivariable table file in csv format.
f_month2day(file): interpolation of daily records from a monthly time series
f_properties(file, gaps = FALSE) displays the properties of a hts time-series.

Hydro-meteo time-series (prefix `h_` or `hs_`)

Data manipulations

h_addna(file), adds records with NA in a time series at given dates. The output file is named with the prefix nap_.
h_changetz, allows to change the timezone of a time series
h_common(files), extracts 2 (or more) time-series on their common period. Generates a file with the prefix co_.
h_condition(files, condition), conditionally extracts a time-series regarding another one. Generates a file with the prefix cd_.
h_cumul(file, start = NA, end = NA), cumulates the values of a time-series. Generates a file with the prefix cu_.
h_nodata(file, threshold=NA, test="=", start=NA, end=NA), replaces values with NA conditionally or in a time interval. Generates a file with the prefix na_.
h_rbind(files, sensor, gap = TRUE), binds 2 time-series on consecutive periods
h_replace(file, old.val, new.val), replaces a value by another. Generates a file with the prefix re_.
h_rollav (file, ti = 7, position = c("central", "right")), computes a rolling average of a daily time-series. Generates a file with the prefix ro_.
h_substitute(files), substitutes the missing values in a series by existing values of another series. Generates a file with the prefix su_.
h_weightedsum(files, weights = NA, constant = 0), makes a weighted sum of time-series. Generates a file with the prefix ws_.

Data critics

h_gaperr(file, nv = 1, itv0 = 43201, df), replaces errors with gaps in a time-series based on neighboring values. Generates a file with the prefix eg_.
h_gapfill(file, npdt), produces a simple gapfilling in a time-series. Generates a file with the prefix gf_.
h_gaprem_itv(file, itv0 = 43201), removes gaps in a time-series with a time interval threshold. Generates a file with the prefix gr_.

Time treatments

hs_tstep(), computes infra-daily data with a fixed time step. Generates a file with the suffix _xxxx, where xxxx is the tst value. It also makes monthly operations, based on a daily time-series. Generates hts files with the suffixes _C, _G or _M and MS Excel files with the prefixes ad_ and cm_.
h_year(file, mhy = 1, op = "M", dig = 1) extracts an annual time series from a daily time series.
h_restrict(file, start=NA, end=NA), restricts a series between 2 dates. Generates hts files with the suffixes re_.
h_season(file, monthstart), produces a seasonal selection. Generates a file with the prefix sx_, where x is 2, 3 or 4.
h_avday computes a one year time series filled with the mean values of each calendar day over an interval longer than 4 years.

Hydro-meteo processes

h_stat_basic(file), gives basic statistics of a time-series
h_adjust(file, time_unit = "year"): adjustment of a time-series to a statistical model. For instance only a linear model is allowed.
h_wl_di(fsq, sta, seni, seno, dstart = NA, dend = NA, dbo = TRUE), computes a discharge time-series from water levels data and calibration curves.
h_rainsnow(fpr, fta, ta0,ta1,sta=NA), shares the solid and liquid precipitations with a temperature criteria. Generates a file with the prefix pr_. Generates a file with the prefix sn_.

Plot functions (prefix `p_` or `ps_`)

General plots

ps_plothts(), plots hts files.

Other plots

p_box_month(file, title = "Title", axeY = "Y-axis", savefig = FALSE, fileo = "plot.png", width = 8, height = 6), plots a boxplot of the 12 months of a time-series.
p_clim(p_clim <- function (files, type="line", hydro.month=1, title="Title", yaxis="Value", y.down=NA, y.up=NA, rpal=FALSE, pal=mapalette, legend.l=NA)), plots climatologies in hydrological year.
p_discalib(fcalib, sen, plotcalib= TRUE, plotdism=TRUE, title="Title", savefig=FALSE, width= 8, height= 6, fout="plot.png", limx =FALSE, limy = FALSE, xinf=NA, xsup=NA, yinf=NA, ysup=NA), plots calibration curves water levels vs discharges.
p_gaps(nbf, title = "Inventory", BW = FALSE, margin = 0.1), plots of data inventory
p_hypso(file, abbrev, prop = FALSE, range=50, fact=5, title="Title", savefig=FALSE, width= 8,height= 6, fileo="plot.png"), plots the hypsometry curve of one or more basins
p_scatter(files, intercept.zero = FALSE, remove.zero = FALSE, lg.axis = c(NA, NA),title = "Title"), plots a scatter plot of 2 or more time-series
p_wind(fsq, sta, swd, swv, ws.int = 0.5, angle = 45, grid.line = 10, type = "default", breaks = 5, offset = 5, paddle = FALSE), plots wind rose

Weather functions (prefix `w_`)

w_atmp_alt (f_atmp, f_temp, alt0 = 0, alt) computes atmospheric pressure, function of altitude.
w_etp (method = c("Turc", "Penman-Monteith", "Priestley-Taylor", "Makkink", "Heargraves-Samani"), freq = c("day", "month"), f_temp, f_relh = NA, f_radg = NA, f_radn = NA, f_atmp = NA, f_wvel = NA, f_tmin = NA, f_tmax = NA, lat = NA, alt = NA, albedo = NA, z = NA), computes the potential evapotranspiration with several methods. Generates a file with the prefix xEtpyy_, where x is J (daily) or M (monthly), and yy is Tu (Turc), PM (Penman-Monteith), PT (Priestley-Taylor), Ma (Makkink) or HS (Heargraves-Samani).
w_temp_alt (file , alt0 = 0, alt, grad = -0.0065) computes temperature, function of altitude.
w_spc2rel_hum ((f_spechum, f_temp, f_atm) converts specific humidity to relative humidity.

Miscellaneous functions (prefix `z_`)

z_coord(ncoord = NA, ccoord = NA, type), converts coordinate from numeric to character and reverse.

Development main references

SQLite management

SQLite, sql database engine, 2017, Tutorial Point, https://www.tutorialspoint.com/sqlite/sqlite_pdf_version.htm
Ripley B., ODBC Connectivity, 2020. https://CRAN.R-project.org/package=RODBC
https://www.sqlite.org/index.html

R coding

Chang W., 2013. R Graphics Cookbook. O’Reilly. 398p. ISBN 978-1-449-31695-2. Actualized on https://r-graphics.org/
Wickham H., 2015. R Packages - Organize, test, document, and share your code. O’Reilly. 190p. ISBN 978-1-491-91059-7. Actualized version on https://r-pkgs.org/
Wickham H., 2021. Mastering Shiny. 348p. ISBN 978-1-492-04738-4. https://mastering-shiny.org/index.html
Wickham H. & Grolemund G., 2017. R for Data Science - Import, tidy, transform, visualize, and model data. O’Reilly. 494p. ISBN 978-1-491-91039-9. https://r4ds.had.co.nz/
https://www.tidyverse.org/

Appendix

Correspondence of the data base tables between Hydraccess and Sqlite htsr

Main tables

--------------------------------------------
  Designation     Hydraccess    Sqlite name 
--------------- -------------- -------------
   Discharge        Debits          DI      
 Precipitation      Pluies          PR      
    Quality        Qualite          QU      
    Sensors        Capteurs         SS      
    Station      Station_Base       ST      
    Weather         Meteo           WE      
  Water level       Cotes           WL      
--------------------------------------------

Other tables

--------------------------------------------------------------
        Designation              Hydraccess       Sqlite name 
--------------------------- -------------------- -------------
           Basin                  Bassins             BA      
     Calibration Date            Etal_Dates           CD      
          Country                Zones_Pays           CO      
 Disch. measur. proceeding     Jaugeages_Dep          DP      
   Discharge measurement         Jaugeages            DM      
      Elevation zero              Zero_NG             EZ      
         Equipment              Equipements           EQ      
           Event                 Evenements           EV      
        Large basis            Bassins_Grands         LB      
  Liquid flow calibration         Etal_HQ             LC      
          Manager              Gestionnaires          MG      
         Nature id              Codes_Nature          NC      
      Operating mode         Modes_Opératoires        OM      
        Origine id.            Codes_Origine          OC      
          Profile               Profils_Data          PF      
       Profile data             Profils_Data          PD      
         Propeller                Helices             PP      
        Quality id.            Codes_Qualite          QC      
          Region                  Regions             RE      
           River                  Rivieres            RV      
       Sensor comm.            Capteurs_Comm          SM      
      Sensor history          Capteurs_HistApp        SH      
         Settings               Parametrage           SE      
        Small basin            Bassins_Petits         SB      
  Solid flow calibration          Etal_HK             SC      
     Station equipment       Stations_Equipment       SQ      
       Station file          Dossiers_Stations        SF      
         Sub-zone                Zone_Sous            SZ      
       Temp_station2           Temp_Stations2         TS      
           Valve                   Vannes             VA      
           Zone                    Zones              ZO      
--------------------------------------------------------------

Fields of the main tables

They correspond to French spelling, which is used in Hydraccess. The compulsory fields are marked with (!).

Stations_Base/ST

Ordre, Type_Station (!), Id_Station (!), Id_Secondaire, Id_Tertiaire, Type_Meteo, Nom, Pays, Zone, SousZone, GrandBassin, Bassin, PetitBassin, Riviere, Gestionnaire, Latitude, Longitude, Altitude, Superficie_bv, Mois_Debut_Hydro, Debut_Activite, Activite, Critere_OuiNon, Critere_Numerique, Critere_Texte, Nom_Observateur, Adresse, Teletransmission, Enregistreur, Fictive, Commentaire, Flag, District, Localite

Capteurs/SS

Type_Station (!), Id_Station (!), Capteur (!), Table (!), Nature, Description, Commentaire, Code_Limni, Principal, Fictif, Maj_Journaliers, Maj_Traduction, Acquisition_Auto, Operationnel, Liste_Inst, Liste_Jour, Liste_Mois, Agregation, Decalage_Temps, Mini, Maxi, Gradient_Maxi, Precision, Decimales, Pente

Cotes/WL, Debits/DI, Meteo/WE, Qualite/QU

Type_Station (!), Id_Station (!), Capteur (!), Table (!), Date (!), Valeur, Origine, Qualite

Pluies/PR

Type_Station (!), Id_Station (!), Capteur (!), Table (!), Date (!), Valeur, Origine, Qualite, Nature

htsr-package

Pierre Chevallier

Content

Origin

Brief history

Hydraccess

Sqlite data base

Principle

Dependencies

Special case of RODBC

File formats

Sqlite data base

Time-series

Gaps

Functions

Short-cut function

Data base functions (prefix d_ or ds_)

Data base tools

Data import/export

Conversion

File functions (prefix f_)

Hydro-meteo time-series (prefix h_ or hs_)

Data manipulations

Data critics

Time treatments

Hydro-meteo processes

Plot functions (prefix p_ or ps_)

General plots

Other plots

Weather functions (prefix w_)

Miscellaneous functions (prefix z_)

Development main references

SQLite management

R coding

Appendix

Correspondence of the data base tables between Hydraccess and Sqlite htsr

Main tables

Other tables

Fields of the main tables

Stations_Base/ST

Capteurs/SS

Cotes/WL, Debits/DI, Meteo/WE, Qualite/QU

Pluies/PR

Data base functions (prefix `d_` or `ds_`)

File functions (prefix `f_`)

Hydro-meteo time-series (prefix `h_` or `hs_`)

Plot functions (prefix `p_` or `ps_`)

Weather functions (prefix `w_`)

Miscellaneous functions (prefix `z_`)