Spark interpreter has z.run(paragraph_id) <https://github.com/apache/zeppelin/blob/master/spark/src/main/java/org/apache/zeppelin/spark/ZeppelinContext.java#L286>. Although it currently supports paragraph in the same notebook, it might be still useful.
Thanks, moon On Thu, Aug 25, 2016 at 3:27 AM Pankaj Singh <[email protected]> wrote: > Thanks Moon for reply. > > https://issues.apache.org/jira/browse/ZEPPELIN-1363 will help to achieve > "input box at the top and bind the remaining paragraphs with it". Running > all the paragraphs is a good solution but i believe we should also provide > something like z.runParagraph(paragraph) > > I will try your suggestions for couple of requirements. > > for "Click visualization element invoke action" agree there should be a > simpler way > > Thanks, > Pankaj > > > On Wed, Aug 24, 2016 at 4:46 PM, moon soo Lee <[email protected]> wrote: > >> Hi Pankaj, >> >> Appreciate for great feedbacks. >> >> 1. What I need is one input box at the top and bind the remaining >> paragraphs with it >> I have created an issue >> https://issues.apache.org/jira/browse/ZEPPELIN-1363. it might be related. >> Zeppelin's interpreter abstraction allows any interpreter creates >> something like Z context. To expose Z context to user in %jdbc interpreter, >> we need proper syntax for it. i.e. %spark interpreter expose Z context as >> scala api. %pyspark interpreter expose Z context as python api. %jdbc >> interpreter expose Z context as ..? >> >> 2. Store query result into a spark data frame >> in Master branch (0.7.0-SNAPSHOT), table type result form any interpreter >> can be accessed by resource pool. You'll get some idea from >> >> https://www.zeppelinhub.com/viewer/notebooks/bm90ZTovL21vb24vTWFjQm9vay82NzcyNjU0NTI3YTY0YTk1YWE4NDhkMGU2Y2QzYTUzMS9ub3RlLmpzb24 >> >> 3. Using basic display system with query result. >> Yes, you can use %html in the cell in table output. Check >> http://zeppelin.apache.org/docs/latest/displaysystem/basicdisplaysystem.html#table >> >> 4. Pulling some paragraphs from different notebooks. >> I know some 3rd party service supports this feature[1], but currently >> Zeppelin doesn't. >> I think simplest way of supporting this(or similar) feature is let >> z.run(PRARAGRPH_ID) api work with paragraphs not only in the same Note but >> also with all other notebooks. >> Then a master notebook can call the other paragraphs using >> z.run(PARAGRPH_ID) and collect graphs. >> >> 5. Authorization in dynamic form. >> Thanks for good suggestion. Do you mind file an issue? >> >> 6. Label overlaps >> Do you mind file an issue for this? >> >> 7. click visualization element invoke action. >> You can create custom visualization in JS inside of Zeppelin notebook and >> watch bind any action using angular display system [2]. >> But it'll be nice if there're some simpler way that can leverage build-in >> visualization. >> >> Thanks, >> moon >> >> [1] http://help.zeppelinhub.com/creating_composers >> [2] >> http://zeppelin.apache.org/docs/0.7.0-SNAPSHOT/displaysystem/back-end-angular.html#watch-unwatch-variables >> >> >> >> On Fri, Aug 19, 2016 at 12:46 AM Pankaj Singh <[email protected]> >> wrote: >> >>> Hi All, >>> >>> Couple of weeks I was experimenting zeppelin to create interactive >>> visualization. Since the cleaned data was available in mysql database in >>> structured format %jdbc was the obvious choice. >>> >>> Since I wanted the dashboard to be interactive I used available dynamic >>> forms available by default. >>> >>> Below are couple of challenges which I faced while using %jdbc >>> interpreter, If it seems valid enhancement I can go and create defects for >>> them >>> >>> >>> >>> 1. *Something similar to z context in %jdbc interpreter* - I have a >>> *input >>> box* which takes number of days as input . I have a number of >>> paragraph which has same *input box.* What I need is one input box >>> at the top and bind the remaining paragraphs with it. ? >>> 1. Can I use something similar to Z context here >>> 2. *Store query result into a spark data frame* : The second >>> requirement is to execute a mysql query using %jdbc and store the result >>> as >>> spark data frame. Basically the requirement is to create multiple temp >>> table after executing a single sql query and create various type of >>> visualization based on those temp queries. >>> 3. *Using Basic Display System with query result* : Can I use %html >>> inside query output . The requirement is – A query output has a column >>> which stores url and I wish to appear as an anchor tag so that it is >>> clickable. >>> 4. *Easy way of pulling some paragraphs from different notebooks* : >>> I have different groups of user which need different type of >>> visualization >>> , the visualization which I want to show can be same/different or mixed. >>> Is it possible to create a master notebook with alls the paragraph and >>> then >>> create notebooks by pulling some paragraph from it. Its not copy but can >>> be >>> thought as reference. >>> 5. *Authorization* - I have used Dynamic form ( text box), Now I >>> want to authorize user to modify the value in text box and execute the >>> paragraph but does not allow him to read/modify the sql query. Currently >>> with the available permission (owner/read/write) its not possible . >>> 6. *Default visualization* – The default visualization bar chart >>> shows the label horizontally. When the number of label increases say to >>> 40-50 or length of text of a label increases I was able to see overlaps >>> in >>> the label text. Is there a way to choose particular portion of the graph >>> and then zoom that portion. >>> 7. *click on a visualization element to invoke any paragraph action* >>> : To make visualization sort of drill down is it possible to >>> invocation a different paragraph by clicking a bar chart. Its like taking >>> input from bar chart as an argument to next paragraph and then execute >>> and >>> transition to that paragraph. >>> >>> Thanks for your inputs and help >>> >>> ~Pankaj >>> >>> >
