Hi Marco, yes this is a good idea. Therefore, we have to extend our model in the data lake sink and store the mapping in the couchDB. We are currently building an API to manage the data within the data lake.
Philipp > On 6. Jul 2020, at 18:14, Marco Heyden <[email protected]> wrote: > > Hi Philipp, > > Maybe an option would be to use a UUID as table name and then create a > mapping from the UUID to the human understandable name of the component. Then > the human could change the name of the component to whatever he wants as we > only need to adapt the mapping UUID → Component Name. > > Best > Marco > Am 6. Juli 2020, 14:58 +0200 schrieb Philipp Zehnder <[email protected]>: >> Hi, >> >> Ok. I agree that a simple default value might be sufficient. >> >> For visualisations on the dashboard the name can be changed directly in the >> dashboard. >> The problem could be with the data lake. Since we use this name for a >> measuerment (table) it must be unique and cannot be changed once it is >> created. >> @Dominik, do you think it is possible to ensure that the value is unique? >> >> Philipp >> >> >>> On 5. Jul 2020, at 20:36, Dominik Riemer <[email protected]> wrote: >>> >>> Hi, >>> >>> that's a good point. In general, we only need these names for the data lake >>> and dashboard sinks to let users distinguish between two visualizations in >>> case multiple dashboards are used per pipeline. >>> So I think using the data source name wouldn't solve the problem as the >>> data source could be split and forwarded to two dashboard sinks (and then >>> the name would be used twice). >>> >>> I personally like the proposal to use increasing numbers as default values >>> and it shouldn't be hard to implement. >>> >>> Dominik >>> >>> On 2020/07/02 14:10:05, Philipp Zehnder <[email protected]> wrote: >>>> Hi Marco, >>>> >>>> welcome to the mailing list. >>>> >>>> It's great that you like StreamPipes. >>>> This is a very good point and I got similar feedback from multiple users >>>> and we should find a solution for that. >>>> >>>> I think the biggest problem is to find meaningful default values, because >>>> when we just use names like Data Lake 1 the user will not know what kind >>>> of data it represents. >>>> >>>> @all: Would it be possible to use the data source name (e.g. Flow Rate >>>> Sensor 2) as a default value? Since the user can do basically anything >>>> within a pipeline I think it is hard to realize this. Any thoughts? >>>> >>>> Philipp >>>> >>>>> On 1. Jul 2020, at 23:16, Marco Heyden <[email protected]> wrote: >>>>> >>>>> Hey everyone, >>>>> >>>>> I played around with streampipes during the last days and overall had a >>>>> great experience. >>>>> However, while setting up the adapters, pipelines, dashboards, data >>>>> sinks, etc. I found it a little annoying to come up with a meaningful >>>>> identifier for each component over and over again as there are no default >>>>> values. For example, I ended up calling the sinks S1, S2, S3 and so on. >>>>> >>>>> I’d like to ask if you had similar experiences and if you have any ideas >>>>> regarding good default values. For now, I suggest simple default values >>>>> such as >>>>> >>>>> • Data Lake 1, 2, 3 ... >>>>> • Dashbord 1, 2, 3 ... >>>>> >>>>> Although they are not really meaningful, prototyping would become more >>>>> convenient. Also I feel like most new users don’t know what the >>>>> identifiers are for. Default values could thus also help users to >>>>> understand why and how they should choose them. >>>>> >>>>> Let me know what you think. >>>>> >>>>> Best >>>>> Marco >>>> >>>> >>
