There are many import tools, and this advice page is intended to help you gain a "forest" perspective. Almost all of the advice contained here is optional, but will save you trouble in the long run.
1. Always run a backupdb before each import. This is critical. Being able to run restoredb to remove incorrect data is 1000% easier than selecting each item and pressing the delete key. Please don't skip this.
2. Set up certain things within Timesheet before you begin importing. Time Entry Screens and Time Periods are the most critical parts of your configuration that you cannot import. It is important that your users get assigned to the correct time entry screen with the correct time period, otherwise you will be undoing this. If you are using Custom Roles then the custom roles should be set up prior to importing your users.
3. Read all of the import file format help documents prior to importing, because you can accomplish many imports in several different ways. You can import your Users and assign them to s in a single file, or you can put the Membership in a separate import. There are several options like that, and the only way to catch them all is to read the file specifications of each of the imports.
4. Jot down a list of the things that you want to import and plan them out in logical order. These should be obvious, like you cannot import accruals until the Users and the s are imported. If you choose to use the Membership import instead of specifying the s in each individual import file, then you must run the Membership import last. Here is the overall recommended order of importing:
User
Membership
Dependencies
Dependency Removal
Membership Removal
Accrual
Memorized Time Records
5. Test, test, and re-test before you enable the scheduled imports. There is no preview mode once you enable the schedule imports. They will put error messages in the debug.log file, but you will receive no notification of that. So if you begin scheduling an import that contains incorrect or mis-formed data then you could set yourself up for several hours or days of bad data in production before you catch the problem. At that point getting the bad data corrected may become a very large task, like manually editing several hundred time records.
6. Don't forget to finish that last 1% of your configuration before you go live. Rules and Policies generally cannot be built and assigned until after the imports are complete, because they depend on so many imported items. Also you will want to review the group-assignments of your users and probably remove them all from the All group. Also you will want to enable some of the extra fields for reporting. And you may have default values in some of the columns that you want to hide or remove. It is always a good idea to test logging in as a few users and reviewing their perspective of the software before taking the software into production.
7. Once you have initially imported all of your information you can make updates by importing again and modifying the records. The modify process works exactly the same as the import process, but the file format is much simpler when modifying. There is only 1 required field when modifying (except in Membership and Dependencies), the first field. So the modification import file can have as few as two columns, the first field and the field that you are modifying. So, for instance, if you did not have everyones' e-mail address when you did your initial import of users, then you could import a file later with just the users' Name (the first field in the user import file) and their E-Mail Address (the 11th field, but the only one you want to modify at this time).
8. As you move towards automatically importing from an external source on a scheduled basis, there are a few things that you should know. The import format is actually much more flexible than is documented in these documents. Generally the only field that has to be in any particular order is the first field, because it is the primary key. Also, some fields that are documented to only accept "Y" will also accept "Yes", "yes", "y", and sometimes "si"--you should test the value(s) that you get from your external system. Also, tab-separating the file isn't the only acceptable delimiter. You can comma-sep. The key is that the header row must use the same separator as the rest of the file. One catch is the few fields that accept multiple values in a single field, like the Groups field. Those are really expecting to be comma-sep, so if you are comma-separating the fields you will have to double-quote the field that has commas within it. This is what Excel does by default. If you hit upon a needed solution that won't work for you please contact the support team at support@journyx.com and the support team will help you get it straightened out. If you are going to build an automatic integration then you really need to invest the time and resources into building a second installation of Timesheet and running your tests there before you put the automatic integration in place.
9. If you ever get stuck, or get an error message that you cannot resolve, please feel free to contact the support team at support@journyx.com and the support team will help you. Please read over the support process and give the support team all of the information that they ask for. The top three causes for people having problems with the import are 1) they have mis-formatted their file somehow, 2) they somehow are running an old version of the tool, and 3) they have funky data in their database already that cannot be overwritten. The support process is designed to catch those three problems quickly. The support process is here.