Should I key-in data in Hydrognomon or in Excel?
Do whatever suits you. We think that keying-in time series in Hydrognomon is easier than keying them in Excel. The reason is that Hydrognomon is specialized and has several facilities (e.g. automatic viewing as a table, quick sums and other statistics, highlighting of too large or small values, and so on). However, it requires you to learn these tricks.
If you choose to key in the data in Excel and then copy/paste it to Hydrognomon, it is best to enter the data in two columns, date(/time) and value, and use yyyy-mm-dd as the date format (yyyy-mm-dd HH:mm if there are times). See the question “How to copy and paste from Excel” for more information.
How do I create a new, empty time series?
Select File, New, and a wizard will appear with three steps:
- General information. While filling in the information in this step is optional, it will really be better for you later to have specified a title, variable, measurement unit, and precision (in many variables, such as temperature and rainfall, the precision is usually one decimal digit and not the default, 2).
- Time step. Select “Time step is defined”, and specify the time series step. If the time stamps don’t all end at exactly the same time but have minor irregularities, uncheck “Time step is strict”.
- Interval. For temperature, select “instantaneous”. For rainfall, select interval and sum. Likewise for other variables. You generally don’t need to select “Advanced settings”.
You can change these properties later by selecting File, Series properties.
How do I key in time series records?
After creating the time series (or opening an existing one), press Ctrl+i. Key in the date and press Enter (it’s way better to forget you have a mouse). Then key in the value and press Enter.
From now on, each time you press Ctrl+i, the date it will propose by default is likely to be the one you want.
What date format should I use?
Because of the ambiguity of several date formats (e.g. it is not clear whether 5/11/2003 means 5 November or 11 May), hydrognomon always displays yyyy/mm/dd (but also accepts yyyy-mm-dd as input). When given dates in other formats it will do its best and will most often parse them correctly, but really, do yourself a favour and always use yyyy-mm-dd.
Can I automatically create many records with their time stamps, so that afterwards I can fill in only the values?
Yes. Press Ctrl+Shift+i. For this to work the time series must already have at least one record.
With Ctrl+Shift+i I can add records at the end or the beginning. Can I add records in the middle?
You can if all the existing time stamps end at the same time. Select Series, Regularize step. Leave the defaults, except that you should blank the field “In new records with dates which did not previously exist raise flag” (the default is DATEINSERT – enter the field and delete its contents with the backspace key). After clicking OK, it will create a new time series which will probably be as you want it.
If the time stamps don’t all end at the same time, you cannot do this. We aren’t certain how you could do it. Maybe you could create a new time series, key in what you want, and then join the two time series together by copying and pasting.
I have a time series with many empty records and I want to set them to zero.
If the time series has only empty records (for example, if it is new), press Ctrl+a to select all records, then press Ctrl+F2 to define a value.
If it has some non-empty, non-zero records, and some empty records, and you want to specify a value for the empty ones, press Ctrl+F9 and specify “Only display records such that the value is null”. Next, press Ctrl+a to select all visible records, then Ctrl+F2 to define a value. Finally, press F9 to toggle the view filter.
Time steps, time stamps, and processing
All time stamps end in 00:00, and I want to change it to 08:00.
Use Tools, Regularize step, and then select “Change time: move each record as it is”.
Or store the time series in a file, close it (Ctrl+w), open it in an editor such as Notepad or Wordpad, and use the editor’s Find & Replace functionality to change all “00:00” to “08:00”.
My daily time series appears to be one day off.
Time series records can refer to an instant or an interval. When we say that the temperature at 15:10 is 19.1°C, we usually mean that this was the temperature at that instance. When a meteorological station set to measure at 10-minute intervals reports that the rainfall at 15:10 was 2 mm, this actually means that this was the rainfall in the interval 15:00–15:10. Time series derived from smaller time steps always refer to intervals; the daily temperature, for example, is usually the average of the temperature during a day.
When we talk about the temperature in 15 July 2016, you might think that this is the average from 2016-07-15 00:00 to 2016-07-16 00:00. However, this is not always the case. For example, in the Hellenic Meteorological Service, the temperature of “15 July 2016” traditionally meant the temperature at the instant 2016-07-15 08:00, because that was the time when the observer visited the station and recorded the measurements. He would also record the rainfall measurement and would empty the rainfall gauge. So the rainfall recorded at 2016-07-15 08:00, or the rainfall of “15 July 2016”, was actually the rainfall from 2016-07-14 08:00 to 2016-07-15 08:00. They are likely to have kept this convention with automatic meteorological stations as well, so the average might well be taken between 2016-07-14 08:00 to 2016-07-15 08:00.
In any case, when the value refers to an interval, the time stamp in Hydrognomon refers to the end of the interval, so a time stamp of 2016-07-16 00:00 refers to day 2016-07-15. The exception is for monthly and annual time series. Since it does seem quite strange that 2016-07-16 00:00 refers to day 2016-07-15, a good compromise is to use the time stamp 2016-07-15 23:59.
So, first, you must be certain about what exactly your time stamps mean. Now, if indeed they are off, there are two ways with which you can fix them:
1. Select Tools, Regularize Step, and then select “Change time: move each record as it is”.
2. Create a new time series and fill it in with empty records and the desired time stamps, then copy and paste the values from the original time series.
I sum a daily time series into a monthly time series, and the summing appears to be offset by one day.
Please see the previous question.
What is the difference between “Regularize step” and “Irregular to strict”?
Regularize step is intended to fix small deviations. For example, a station might record at :10, :20, :30, and so on, and one day it might continue with :42, :52, etc. Or a human observer might record the measurements every day at 08:00, but one day he might oversleep and record them at 10:00. Use the Regularize step function to fix these.
Irregular to strict is intended to normalize time series with completely irregular time step or a very small step (e.g. two-minute or three-minute). It can provide a solution even in large deviations (in which case Regularize step will leave empty records) but must be used carefully.
How can I multiply all values of a time series with a number, or add a constant number to all values?
Use Tools, Linear combinations.
How can I import time series from a text file?
Create a new, empty time series in Hydrognomon, then use File, Import from text file.
Another way is to use Excel’s wizard to open the text file in Excel, then copy and paste the time series to Hydrognomon.
How can I copy and paste from Excel?
Case 1: Data is in two columns, timestamp and value. The date must be in yyyy-mm-dd format (Hydrognomon understands many date formats, but because some formats are ambiguous it’s best to help it). In Excel, select the column with the dates, Format cells, and select custom format, yyyy-mm-dd, or, if you also want time, yyyy-mm-dd HH:mm. Then, in Excel, select the time series and copy it (to do it fast, select the first timestamp cell, and then press Shift+Left, Ctrl+Shift+Down, Ctrl+C). In Hydrognomon, press Ctrl+V (you must have already created the time series).
Case 2: Suitable table. You have a monthly time series in a 12-column table, or a daily time series in a 31-row, 12-column table, or an hourly time series in a 24-column, 31-row table. (Note: It’s better to not have such stuff in Excel because it’s very inconvenient; it’s much easier to use Hydrognomon for that.) In Excel, select only the internal part of the table (without the column and row headings) and copy. Go to Hydrognomon and select Edit, Paste monthly table or Paste daily table, depending on the case. If you are going to do this many times you need to learn the keyboard shortcuts that are written on the menu.
Case 3: Unsuitable table. You have a daily time series in a 31-column table, but Hydrognomon needs it in 31 rows. You need to copy it in Excel, and then, again in Excel, in a different area or sheet, Edit, Paste Special, Transpose. In addition, Hydrognomon needs a separate paste for each year; you can’t paste a long table.