Data sources
A data source says where a file sits and how TRADElube signs in there. You reach the list through the Settings navigation entry and the Data Sources button.
It deliberately sits next to the channel, not inside it. A supplier often delivers more than one file, and the same sign-in is then needed by several tasks. Were the login details part of the channel, you would have to maintain them several times and follow up in several places after a password change.
Overview / list
Three buttons sit above the list. New opens a menu, Edit opens the selected data source, Delete removes it after a confirmation. The latter two require a selected row.
Better not to delete a data source that an endpoint still uses. The task behind it will no longer find its file.
The list shows two columns:
| Column | Content |
|---|---|
| Name | The name you gave it. It later appears in the endpoint's pick list, so it should say whose server this is |
| Type | FTP, SFTP or HTTP |
Sorting is fixed to Name, ascending. There is no Filter field, no Filter By button and no pager here: a client has a handful of data sources, not thousands.
The type is chosen when creating and is fixed afterwards
A click on New opens a menu with the three source types.
The entry decides which fields the new data source carries and cannot be changed later. Where a supplier's FTP server becomes an SFTP server, you create a new data source and switch the endpoint over to it.
Detail view
A data source has two sections, and the second one looks different per type.
Common
| Field | Meaning | Note |
|---|---|---|
| Name | The designation in the list and in the endpoint's pick list | Free text |
| (no caption) | The type, to the right of the name | Text only, not a field. It is fixed from creation |
Login details for FTP and SFTP
The Connect button above the fields opens the connection right away and reports Connection successful or Connection failed. A click on the message shows what the server answered. Use it before setting a task on the source: a typo in the directory shows up here in two seconds, otherwise only during the nightly execution.
| Field | Meaning | Note |
|---|---|---|
| Host | The machine name or address of the server | Without the protocol |
| Port | The port the server listens on | Preset to 21 for FTP and to 22 for SFTP |
| Directory | The directory TRADElube works in | Empty means the login's start directory. Leading and trailing slashes are removed by TRADElube itself |
| User Name | The account TRADElube signs in with | |
| Password | The matching password | Stored encrypted and never shown in clear text again |
Both types carry the same fields, they differ only in the preset port and in the protocol. SFTP transfers encrypted, FTP does not. Where the supplier offers both, SFTP is the right choice.
Login details for HTTP
There is no Connect button here.
| Field | Meaning | Note |
|---|---|---|
| Url | The base address the endpoint's route pattern is appended to | With the protocol, https:// for instance |
| Authentication | Whether and how TRADElube signs in | Empty means without a sign-in. The only kind offered is Basic Authentication |
| User Name | The account for the sign-in | Only appears once an authentication is chosen |
| Password | The matching password | Only appears once an authentication is chosen, stored encrypted |
An HTTP data source can only read. TRADElube fetches a file over it but places none there, and moves and deletes nothing either. For an export you therefore need FTP or SFTP.
The endpoint: where a data source takes effect
A data source alone fetches nothing. It is used as soon as an endpoint names it, and the endpoint sits on the task in the Configuration section. How an endpoint is created and what happens without one is described under the custom channel. Here are the fields it brings along per source type.
A click on the row expands the fields below it. Which ones these are is decided by the chosen data source.
For FTP and SFTP:
| Field | Meaning | Note |
|---|---|---|
| File Pattern | Which files in the directory the task picks up | Preset to {Name}-{Exclude:processed}.ext |
| Processed File Pattern | What the file is called after the task has read it | Preset to {Name}-processed.ext |
| Auto Cleanup | Whether TRADElube deletes old processed files by itself | Without the tick they stay until somebody removes them |
| Retention Period | After how many days a processed file is deleted | Preset to 7. Only appears with the box ticked |
In the pattern {Name} stands for the name of the data format, * for any sequence of characters, and {Exclude:processed} excludes files whose name continues with "processed" at that position.
This pair of two patterns is what keeps an execution from reading the same file twice: whatever is processed gets renamed and thereby drops out of the search pattern. The file still stays on the server, and you can look up what the supplier actually delivered.
For HTTP: a single field, Route Pattern, preset to {Name}. It is appended to the data source's Url and together they form the address TRADElube calls.
Further reading
- Setting up data sources: the way from the supplier's login details to a running collection
- Custom: the channel that exchanges files, with data formats and endpoints
- Settings: which button is covered on which page