Skip to main content

Hydrological and meteorological timeseries

Project description

htimeseries - Hydrological and meteorological time series Coverage Updates

This module provides the HTimeseries class, which is a layer on top of pandas, offering a little more functionality.


from htimeseries import HTimeseries

ts = HTimeseries()

This creates a HTimeseries object, whose data attribute is a pandas time series or dataframe with a datetime index. Besides data, it can have other attributes which serve as the time series’ metadata. There are also several utility methods described below.

HTimeseries objects

HTimeseries(data=None, format=None, start_date=None, end_date=None, default_tzinfo=None)

Creates a HTimeseries object. data can be a pandas time series or dataframe indexed by datetime or a file-like object. If it is a pandas object, it becomes the value of the data attribute and the rest of the keyword arguments are ignored.

The data attribute should be a dataframe with two columns (besides date): value and flags. However, in this version, HTimeseries does not enforce that. A good idea is to create an empty HTimeseries object with HTimeseries(), and then proceed to fill in its data attribute. This ensures that the dataframe will have the right columns and dtypes.

If the data argument is a filelike object, the time series is read from it. There must be no newline translation in data (open it with open(..., newline='\n'). If start_date and end_date are specified, it skips rows outside the range.

The contents of the filelike object can be in text format or file format (see “formats” below). This is usually auto-detected, but a specific format can be specified with the format parameter. If reading in text format, the returned object just has the data attribute set. If reading in file format , the returned object also has attributes unit, title, comment, timezone, time_step, interval_type, variable, precision and location. For the meaning of these attributes, see section “File format” below.

These attributes are purely informational. In particular, time_step and the other time-step-related attributes don’t necessarily mean that the pandas object will have a related time step (also called “frequency”). In fact, raw time series may be irregular but actually have a time step. For example, a ten-minute time series might end in :10, :20, etc., but at some point there might be an irregularity and it could continue with :31, :41, etc. Strictly speaking, such a time series has an irregular step. However, when stored in a database, specifying that its time step is ten minutes (because that’s what it is, ten minutes with irregularities) can help people who browse or search the database contents.

The location attribute is a dictionary that has items abscissa, ordinate, srid, altitude, and asrid.

The timestamps of the data attribute are aware. If HTimeseries() is called with a dataframe, it should have an aware datetime index. If it is called with a filelike object whose contents are in file format and contain the timezone header, that time zone is used to interpret the file’s timestamps. If it is called with a filelike object that does not contain a timezone header, default_tzinfo is used. In the latter case, if default_tzinfo is None or unspecified, an exception is raised. Creating an empty HTimeseries object without specifying default_tzinfo (e.g. with HTimeseries()) assumes default_tzinfo=ZoneInfo("UTC").

.write(f, format=HTimeseries.TEXT, version=None)

Writes the time series to filelike object f. In accordance with the formats described below, time series are written using the CR-LF sequence to terminate lines. Care should be taken that f, or any subsequent operations on f, do not perform text translation; otherwise it may result in lines being terminated with CR-CR-LF. If f is a file, it should have been opened with newline=”n”.

version is ignored unless format=HTimeseries.FILE. The default version is latest.

While writing, the value of the precision attribute is taken into account.

TzinfoFromString objects

from htimeseries imort TzinfoFromString

atzinfo = TzinfoFromString("EET (UTC+0200)")

TzinfoFromString is a utility that creates and returns a tzinfo object from a string formatted as “+0000” or as “XXX (+0000)” or as “XXX (UTC+0000)” (TzinfoFromString is actually a tzinfo subclass). Its purpose is to read the contents of the timezone parameter of the file format (described below).


There are two formats: the text format is generic text format, without metadata; the file format is like the text format, but additionally contains headers with metadata.

Text format

The text format for a time series is us-ascii, one line per record, like this:

2006-12-23 18:34,18.2,RANGE

The three fields are comma-separated and must always exist. In the date field, the time may be missing. The character that separates the date from the time may be either a space or a lower case t, or a capital T (this module produces text format using a space as date separator, but can read text format that uses t or T). The second field always uses a dot as the decimal separator and may be empty. The third field is usually empty but may contain a list of space-separated flags. The line separator should be the CR-LF sequence used in MS-DOS and Windows systems. Code that produces text format should always use CR-LF to end lines, but code that reads text format should be able to also read lines that end in LF only, as well as CR-CR-LF (for reasons explained in the write() function above).

In order to improve performance in file writes, the maximum length of each time series record line is limited to 255 characters.

Flags should be encoded in ASCII; there must be no characters with code greater than 127.

File format

The file format is like this:

Title=My timeseries

2006-12-23 18:34,18.2,RANGE
2006-12-23 18:44,18.3,

In other words, the file format consists of a header that specifies parameters in the form Parameter=Value, followed by a blank line, followed by the timeseries in text format. The same conventions for line terminators apply here as for the text format. The encoding of the header section is UTF-8.

Client and server software should recognize UTF-8 files with or without UTF-8 BOM (Byte Order Mark) in the begining of file. Writes may or may not include the BOM, according OS. (Usually Windows software attaches the BOM at the beginning of the file).

Parameter names are case insensitive. There may be white space on either side of the equal sign, which is ignored. Trailing white space on the line is also ignored. A second equal sign is considered to be part of the value. The value cannot contain a newline, but there is a way to have multi-lined parameters explained in the Comment parameter below. All parameters except Version are optional: either the value can be blank or the entire Parameter=Value can be missing; the only exception is the Comment parameter.

The parameters available are:


There are four versions:

  • Version 1 files are long obsolete. They did not have a header section.

  • Version 2 files must have Version=2 as the first line of the file. All other parameters are optional. The file may not contain unrecognized parameters; software reading files with unrecognized parameters may raise an error.

  • Version 3 files do not have the Version parameter. At least one of the other parameters must be present. Unrecognized parameters are ignored when reading. The old deprecated parameter names Nominal_offset and Actual_offset are used instead of the newer (but also deprecated) ones Timestamp_rounding and Timestamp_offset.

  • Version 4 files are the same as Version 3, except for the names of the parameters Timestamp_rounding and Timestamp_offset.

  • Version 5 files are the same as Version 4, except that Timestamp_rounding and Timestamp_offset do not exist, and Time_step is in a different format (see below).


A symbol for the measurement unit, like °C or mm.


The number of records in the time series. If present, it need not be exact; it can be an estimate. Its primary purpose is to enable progress indicators in software that takes time to read large time series files. In order to determine the actual number of records, the records need to be counted.


A title for the time series.


A multiline comment for the time series. Multiline comments are stored by specifying multiple adjacent Comment parameters, like this:

Comment=This timeseries is extremely important
Comment=because the comment that describes it
Comment=spans five lines.
Comment=These five lines form two paragraphs.

The Comment parameter is the only parameter where a blank value is significant and indicates an empty line, as can be seen in the example above.


The time zone of the timestamps, in the format {XXX} (UTC{+HHmm}), where XXX is a time zone name and +HHmm is the offset from UTC. Examples are EET (UTC+0200) and VST (UTC-0430).

The TzinfoFromString utility (described above) can be used to convert this string to a tzinfo object.


In version 5, a pandas “frequency” string such as 10min (10 minutes), H (hour), or 2M (two months). If missing or empty, the time series is without time step.

Up to version 4, a comma-separated pair of integers; the number of minutes and months in the time step (one of the two must be zero).

When reading from version 4 or earlier, the pair of integers is automatically converted to a pandas “frequency” string, so the time_step attribute of an HTimeseries object is always a pandas “frequency” string. Likewise, when writing to a version 4 or earlier file, the pandas “frequency” string is automatically converted to the pair of integers.


Deprecated. It might be found in old files, Version 4 or earlier, but htimeseries will ignore it when reading and will never write it.

A comma-separated pair of integers indicating the number of minutes and months that must be added to a round timestamp to get to the nominal timestamp. For example, if an hourly time series has timestamps that end in :13, such as 01:13, 02:13, etc., then its rounding is 13 minutes, 0 months, i.e., (13, 0). Monthly time series normally have a nominal timestamp of (0, 0), the timestamps usually being of the form 2008-02-01 00:00, meaning “February 2008” and usually rendered by application software as “Feb 2008” or “2008-02”. Annual timestamps have a nominal timestamp which normally has 0 minutes, but may have nonzero months; for example, a common rounding in Greece is 9 months (0=January), which means that an annual timestamp is of the form 2008-10-01 00:00, normally rendered by application software as 2008-2009, and denoting the hydrological year 2008-2009.

timestamp_rounding may be None, meaning that the timestamps can be irregular.

Timestamp_rounding is named differently in older versions. See the Version parameter above for more information.


Deprecated. It might be found in old files, Version 4 or earlier, but htimeseries will ignore it when reading and will never write it.

A comma-separated pair of integers indicating the number of minutes and months that must be added to the nominal timestamp to get to the actual timestamp. The timestamp offset for small time steps, such as up to daily, is usually zero, except if the nominal timestamp is the beginning of an interval, in which case the timestamp offset is equal to the length of the time step, so that the actual timestamp is the end of the interval. For monthly and annual time steps, the timestamp offset is usually 1 and 12 months respectively. For a monthly time series, a timestamp offset of (-475, 1) means that 2003-11-01 00:00 (often rendered as 2003-11) denotes the interval 2003-10-31 18:05 to 2003-11-30 18:05.

Timestamp_offset is named differently in older versions. See the Version parameter above for more information.


Deprecated. Has one of the values sum, average, maximum, minimum, and vector_average. If absent it means that the time series values are instantaneous, they do not refer to intervals.


A textual description of the variable, such as Temperature or Precipitation.


The precision of the time series values, in number of decimal digits after the decimal separator. It can be negative; for example, a precision of -2 indicates values accurate to the hundred, such as 100, 200, 300 etc.

Location, Altitude

(Versions 3 and later.) Location is three numbers, space-separated: abscissa, ordinate, and EPSG SRID. Altitude is one or two space-separated numbers: the altitude and the EPSG SRID for altitude. The altitude SRID may be omitted.


Htimeseries is

Copyright (C) 2013-2014 TEI of Epirus
Copyright (C) 2018-2021 National Technical University of Athens
Copyright (C) 2018-2021 Institute of Communications and Computer Systems
Copyright (C) 2022 IRMASYS P.C.

Htimeseries is free software: you can redistribute it and/or modify it under the terms of the GNU Affero General Public License, as published by the Free Software Foundation; either version 3 of the License, or (at your option) any later version.

The software is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the licenses for more details.

You should have received a copy of the license along with this program. If not, see

Htimeseries was funded by several organizations:


6.0.3 (2024-03-21)

  • Fixed error with unspecified time zone in HTimeseries objects created empty. Now creating an empty object with HTimeseries() assumes default_tzinfo=ZoneInfo("UTC").

6.0.2 (2023-12-22)

  • Fixed crash when reading csv with aware timestamps.

6.0.1 (2023-12-19)

  • Fixed crash when reading csv with only a date column.

6.0.0 (2023-12-14)

  • Python 3.9 is now required (the reason for this is that backports.zoneinfo is behaving differently from zoneinfo).

  • Only pandas>=1.5,<2 is now supported. Pandas<1.5 probably did not work in 5.0.0 either, but it had not been discovered. Pandas>=2 handles ambiguous times differently and is therefore still unsupported.

  • Increased CSV reading speed by two orders of magnitude.

5.0.0 (2023-11-21)

  • Removed compatibility with pandas<1. pandas 1 and 2 are now supported. This helps avoid an error when reading a data file spanning a time interval that contains a change to or from DST.

4.0.2 (2022-11-27)

4.0.1 (2022-11-27)

  • Fixed packaging error where it didn’t install dependency backports.zoneinfo on Python<3.9.

4.0.0 (2022-11-22)

  • The timestamps of the dataframe index are now aware. When initializing a HTimeseries object with a dataframe, the dataframe index timestamps must be aware. When initializing with a filelike object, either the filelike object must contain a timezone header or the new default_tzinfo parameter must be specified.

3.1.1 (2020-12-29)

  • Fixed incompatibility with pandas 1.2.

3.1.0 (2020-12-17)

  • When reading a time series from a file-like object, it is now permitted for it to not have a flags column. In this case, an empty flags column is automatically added to the HTimeseries object.

3.0.0 (2020-02-23)

  • Only Python>=3.7 is now supported.

  • When reading or writing a time series, it now checks that there are no duplicate timestamps and raises an exception if there are.

2.0.5 (2020-01-15)

  • Fix pandas dependency to use pandas<1, so that Python 3.5 compatibility is kept.

2.0.4 (2020-01-08)

  • Fixed crash when saving in version 2 and the time step was a mere “M” or “Y” without multiplier.

2.0.3 (2020-01-05)

  • Default version when writing file is now latest.

2.0.1 (2020-01-04)

  • Fixed error when the time step was empty.

2.0.0 (2020-01-04)

  • Changed the way the time step is specified. Instead of “minutes,months”, it is now a pandas “frequency” offset specification such as “5min” or “3M”.

  • The timestamp_offset and timestamp_rounding parameters have been abolished.

1.1.2 (2019-07-18)

  • Fixed some altitude-related bugs: 1) It would crash when trying to read a file that specified altitude but not location; 2) it wouldn’t write altitude to the file it the altitude was zero.

1.1.1 (2019-06-12)

  • Fixed crash when Timestamp_rounding=None or Timestamp_offset=None.

1.1.0 (2019-06-08)

  • Added TzinfoFromString utility (moved in here from pthelma).

1.0.1 (2019-06-06)

  • Fixed error in the README (which prevented 1.0.0 from being uploaded to PyPi).

1.0.0 (2019-06-06)

  • API change: .read() is gone, now we use a single overloaded constructor; either HTimeseries() or HTimeseries(dataframe) or HTimeseries(filelike).

  • The columns and dtypes of .data are now standardized and properly created even for empty objects (created with HTimeseries()).

0.2.0 (2019-04-09)

  • Auto detect format when reading a file

0.1.0 (2019-01-14)

  • Initial release

Project details

Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

htimeseries-6.0.3.tar.gz (26.4 kB view hashes)

Uploaded Source

Built Distribution

htimeseries-6.0.3-py2.py3-none-any.whl (14.9 kB view hashes)

Uploaded Python 2 Python 3

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page