Convert Epoch to Date with select value - date

I'm trying to convert a epoch timecode to a date in Pentaho Spoon. I use an input text file to extract fields from. I want to export the fields in a database but there is this timestamp field that contains epoch timestamps like this "1480017396", the datatype is set as an integer and the field is named timestamp. I want to convert with it with Select value.
So I go to the next step and use the select value option to select the field and change the datatype to Date with a format of dd/MM/yyyy the result gives me all kinds of dates in 18-01-1970 range. I tried everything (Different formats etc.) but I just can't seem to solve it.
Any guesses? Image of output

The time in epoch is in miliseconds, not seconds, so, take your number, multiply by 1000, and turn to date.
See that if you divide, the date goes back a few ... and multiply it you get the correct date because of the timestamp.

Related

Copying timestamp format from avro to redshift

I am trying to copy an avro file to redshift using the COPY command. The file has a column that is of the type:
{'name': 'timestamp',
'type': ['null', {'logicalType': 'timestamp-millis', 'type': 'long'}]}],
Redshift variable type: "timestamp" timestamptz
When I run the following command copy if fails:
COPY table_name
from 'fil_path.avro'
iam_role 'the_role'
FORMAT AS avro 'auto'
raw field value: 1581306474335
Invalid timestamp format or value [YYYY-MM-DD HH24:MI:SSOF]
However If I add the following line It works:
timeformat 'epochmillisecs'
I tried to put my timestamp in microseconds which should be the base supported epoch resolution but it fails as well, and didn't find an appropriate name (epochmicrosecs didn't seem to do the job).
My question is why is it so?
Furthermore I have another field that is causing some problem. A date field which apparently is saved as a number of days in the avro file (7305) that gives the following error:
Redshift variable type: "birthdate" date
avro: 'date_of_birth', 'type': ['null', {'type': 'int', 'logicalType': 'date'}]}
Invalid Date Format - length must be 10 or more
Firstly, about the Time Format:
As Docs states:
COPY command attempts to implicitly convert the strings in the source data to the data type of the target column. If you need to specify a conversion that is different from the default behavior, or if the default conversion results in errors, you can manage data conversions by specifying the following parameters.
First Solution:
Redshift Doesn't Recognize epoch time by default to be able to convert it to the format of TimeStamp as a result it can't extract year, month, day..etc from the epoch time to put them in the TimeStamp Format, as stated by the Docs:
If your source data is represented as epoch time, that is the number of seconds or milliseconds since January 1, 1970, 00:00:00 UTC, specify 'epochsecs' or 'epochmillisecs'.
This is the supported Formats that Redshift can convert Using automatic recognition.
TimeStamp needs the format to be as YYYYMMDD HHMISS = 19960108 040809 to be able to extract it right, that's what the error state Invalid timestamp format or value [YYYY-MM-DD HH24:MI:SSOF], while epoch time format is just seconds or milliseconds since January 1, 1970 that it doesn't understand how to extract it's values from.
microseconds isn't supported as a parameter for TIMEFORMAT in Redshift.
Second Solution:
You won't need to pass TIMEFORMAT to the COPY command, but you will insert epoch time in your staging tables as VARCHAR or TEXT.
Then, when inserting epoch time from your staging tables into the schema tables convert it like this: TIMESTAMP 'epoch' + epoch_time/1000 * interval '1 second' AS time
Secondly, about date field:
DATE data type is specified as Calendar date (year, month, day) as stated by the Docs, As a result it can't be the number of days or be less than 10 characters in length (as 2021-03-04) and that's what the error tell us Invalid Date Format - length must be 10 or more.
The solution for Date field:
You need to do a work-around, by passing the number of days as a VARCHAR or text to your staging tables.
When loading the schema tables from the staging tables, apply Data cleaning by convert number of days to a DATE using TOCHAR: TO_DATE(TO_CHAR(number of days, '9999-99-99'),'YYYY-MM-DD')
As a result, number of days will be a valid DATE in your schema tables.

How do I convert a Julian date stored as a double-precision value to a timestamp with at least one-minute resolution?

I exported data from an SQLite table to a CSV file. The data includes a timestamp with at least one-minute resolution: "2019-11-15 01:30:06". The data is actually stored as a Julian date, in this case 2458802.35424295. I imported the data into a double-precision field. I need to convert that number into a timestamp with time zone. I tried casting the double-precision number to text and then using to_timestamp(), but that appears to work only with integer days. I can get a timestamp, but it is always at midnight of the correct date. I tried using to_timestamp() passing in my number, but that returns an epoch (number of milliseconds since 1/1/1970).
I could try to take the fractional part of my Julian date value, calculate the number of milliseconds since midnight that represents, use the to_timestamp(text,text) method to get the date I need, and then add the epoch since midnight to that date. But that's awfully cumbersome. Isn't there a better way?
I'm using PostgreSQL 9.3.
NOTE: The simple answer to my problem, which occured to me just before I clicked the Post button, is to export the data in the form I want, using SQLite's datetime() function to convert the number to a date string during export. But I remain curious. I would have thought there would be a standard way to do this conversion.

postgreSQL increment number in output

I am extracting three values (server, region, max(date)) from my postgresql> But I want to extract an additional 4th field which should be the numerical addition of 1 to 3rd field. I am unable to use date add function as in the database date field is defined as an integer.
date type in DB
date|integer|not null
tried using cast and date add function
MAX(s.date)::date + cast('1 day' as interval)
Error Received
ERROR: cannot cast type integer to date
Required output
select server, region, max(alarm_date), next date from table .....
testserver, europe, 20190901, 20190902
testserver2, europe, 20191001, 20191002
next date value should be the addition to alarm_date
To convert an integer like 20190901 to a date, use something like
to_date(CAST(s.date AS text), 'YYYYMMDD')
It is a bad idea to store dates as integers like that. Using the date data type will prevent corrupted data from entering the database, and it will make all operations natural.
First solution that came to my mind:
select (20190901::varchar)::date + 1
Which output 2019-09-02 as type date.
Other solutions can be found here.

How to add 13min's time in Time Field(not date time field) Crystal reports

I tried using Dateadd('n',13,{fieldname}).
but it throws an error not a valid date time field cause it is time field.
is there any way to convert it to datetime and add the value and revert it into time field?
Thanks.
Try this formula:
DateAdd('n',13,DateTime(CurrentDate, {fieldname}))
The DateTime(date, time) function will create a DateTime value that works in the DateAdd() function. This will only work if the {fieldname} is a Time data type though.
If {fieldname} is a string, you will need to convert it to a Time data type first using the Time(time) function.
The formula I suggested above will append your time value to today's date. You will then need to format the DateTime value it returns to only display the time value. This can be done by right clicking the field in your crystal report and clicking Format Field and setting the Style on the Date and Time tab.

Date Column Split in Talend

So I have one big file (13 million rows) and date formatted as:
2009-04-08T01:57:47Z. Now I would like to split it into 2 columns now,
one with just date as dd-MM-yyyy and other with time only hh:MM.
How do I do it?
You can simply use tMap and parseDate/formatDate to do what you want. It is neither necessary nor recommended to implement your own date parsing logic with regexes.
First of all, parse the timestamp using the format yyyy-MM-dd'T'HH:mm:ss'Z'. Then you can use the parsed Date to output the formatted date and time information you want:
dd-MM-yyyy for the date
HH:mm for the time (Note: you mixed up the case in your question, MM stands for the month)
If you put that logic into a tMap:
you will get the following:
Input:
timestamp 2009-04-08T01:57:47Z
Output:
date 08-04-2009
time 01:57
NOTE
Note that when you parse the timestamp with the mentioned format string (yyyy-MM-dd'T'HH:mm:ss'Z'), the time zone information is not parsed (having 'Z' as a literal). Since many applications do not properly set the time zone information anyway but always use 'Z' instead, so this can be safely ignored in most cases.
If you need proper time zone handling and by any chance are able to use Java 7, you may use yyyy-MM-dd'T'HH:mm:ssXXX instead to parse your timestamp.
I'm guessing Talend is falling over on the T and Z part of your date time stamp but this is easily resolved.
As your date time stamp is in a regular pattern we can easily extract the date and time from it with a tExtractRegexFields component.
You'll want to use "^([0-9]{4}-[0-9]{2}-[0-9]{2})T([0-9]{2}:[0-9]{2}):[0-9]{2}Z" as your regex which will capture the date in yyyy-MM-dd format and the time as mm:HH (you'll want to replace the date time field with a date field and a time field in the schema).
Then to format your date to your required format you'll want to use a tMap and use TalendDate.formatDate("dd-MM-yyyy",TalendDate.parseDate("yyyy-MM-dd",row7.date)) to return a string in the dd-MM-yyyy format.