Microsoft 365 Community Graph, 9 Jul 2020

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: Microsoft Graph community call-July 2020

Description

Hosted by Jeremy Thake, July’s call included what’s next in the Microsoft Search API, new UI for the changelog and updates to Excel in Microsoft Graph plus information about the new TAP program for Microsoft Graph.

Speakers: Jeremy Thake, Nicolas Moreau, Molly Shove, Phoebe Yuan

Resources:
Microsoft Search API https://aka.ms/AA8x21a
Changelog updates https://aka.ms/AA8xna5
Excel in Microsoft Graph https://aka.ms/AA8x21h
TAP program submission https://aka.ms/GraphTAPForm

Stay connected with us on Twitter https://twitter.com/microsoft365dev
Next call is August 2nd at 8:00am PT https://aka.ms/microsoftgraphcall

A

Thank you, everyone for your time today, Jeremy thank I'm gonna run the Microsoft graph Kimiko this month. There are a bunch of other calls that will show the list of the end of this session, and we've got an action-packed one this week from various different workload, PMS across the Microsoft graph and a nice edition, with some showcasing of what we're going to be doing with a change log by Molly Chauvin, our team, which is really exciting, so we're gonna, get straight into it. Nikolas is going to go. First.

A

Nikolas has been on before to talk about search api's. This is something that we have announced a while ago now and have a lot of workloads already in it, and he's gonna be sharing that update and as I mentioned, Molly's gonna be talking about the change log, and then we have Phoebe who's gonna, be sharing some updates on the Excel, API or Microsoft graph as well.

A

If you do have a question, please feel free to put your hand up we'll, be monitoring in the background in there or you can post messages in the chat window as well: Eva, Brian, rye or kind of proposed to speak at the appropriate time and get you to ask the question. We can Mike open if you Pete hands up or in the text messages so without further ado I'll hand over to Nicholas who talked a little bit about the Search API great.

B

Thank you everyone, so for me, it's Monday because I'm back from a long weekend, so that was a good start. Okay, so my name is Nicola. Morrow I have the opportunity to talk in this forum, I think about six or nine months back so I'm part of a team who is building the source platform for Microsoft 365. So primarily, our team is running. Outlook search is running, search upon search, team search and many other workload.

B

So what I'm gonna talk about here is what we are planning to add in the Microsoft Search API for in the next six month. Okay, so just a first recap of what we've announced that ignites last year, so for those who are not familiar with so Microsoft search is the experiences where in Microsoft, if I hope you can search across all your content. You can see this as kind of the successor of Chapman search, so Microsoft as a wider scope. You can search in sharpen pencils, research in teams.

B

We've also announced connectors, which is the ability to ingest external content into Microsoft, 365 and surface this content into SharePoint experience. We are custom verticals, so this is what we this is on the on the end-user visible part, and so our team is building the API that is being used by SharePoint, which is kind of an internal idea that is not exposed publicly and part of that. What we wanted to do is to expose the ability in the Microsoft graph for third-party developers to build similar experience.

B

So one can scenario: if you have, your company has a custom app on mobile, that is providing access to a bunch of content to your employee. Maybe you want to integrate search into this mobile app and if you don't want to use the SharePoint mobile app, because your company is building something very specific, then you need in a search endpoint, and so the Microsoft Search API in the graph will provide you this ability what we've released at ignite. Just a quick recap on the right-hand side.

B

You can see the type of workload that were exposed so email and calendar. Where are the first two things we exposed for email and calendar I think there is a little bit of redundancy with some already available dollar search capabilities on email in calendar, so I won't I. Won't let him do that, but an ignite. We release the ability to search in a pond one drive as well as connectors, which was really one of the key new thing. This is from the. What is the?

B

What is the data available today that you can reason and search on from such capabilities so from or from a search feature perspective? What you can do today? You can scope your search by entity, so today we only support scoping by a single entity, so you can either search files. You can either such external items message. We don't have the ability to do so across multiple type of content and I will talk a lit about that later. The the results, the search results that come back in the workload are actually graph entities.

B

They look very similar to drive items actually in this slide. Here you can see the drive atom. On the right hand, side is actually the model that, if you have today in a canvas that is manipulating drive atoms, so you see you have most of the properties that will drive item, as one caveat here is search is not search, doesn't have access to the full set of data, so we have a bunch of property in the search index. But if, for example, you want to access the preview RL today, the search index doesn't have that.

B

So if you wanted to have access to preview RL man, you would have to do an extra lookup based on them on the idea of the items to get the items to retrieve them. The thumbnail URL so yeah. It's important to note that the graph the search index doesn't have access to all the property, and then the last capability was yes, the ability to run kql queries, um so you would be able to scoop grace such as file type conducts and then whatever search term.

B

So this is what we released, and what I want to talk about a little bit later is what we plan to be added before that. What was the thought process to prioritize? What we wanted to add, so the the key feedback we've got was mostly from SharePoint developers and the to me piece of feedback was a.

B

We don't add all the search capabilities that Japan has and there are some gaps and so important gaps, and that is one of them that we are covering in our next release and the second part is today we were only surfacing drive items and a lot of people are asking for what about site. What about list item? What about list? What about drives so we're gonna, bring all that content that is available in SharePoint and then and the Terr feedback we've received and a really appreciating.

B

There was a lot of engagement on github, so I'm kind of new to get them and I really like, because I've seen people, maybe some people in that core here. Finding some issues in data that have helped us catch some rushon that we have had in our API. So that's the side effect of being better, and this has really really helped us catch some regression.

B

So one of the things that we also want to improve is to make sure that we catch these regressions so that our ABI is consistent and that you can run your app on top of it. So these are the three first pieces of feedback that are very sharp on centric and there was another piece of feedback which is we need more content in search. So today, when you go into especially when you go into Microsoft search in Bing, you can see a lot of many other things.

B

You can see team's conversation, you can see bookmarks answers, so this is something that is currently in planning. We are still figuring out how to prioritize in sequence that work, so this is not coming yet. Unfortunately, so it seems, bookmark answer is not yet coming. So what is coming by the end of this year and makes expect to release some of this stuff by ignite 2020? So the first thing is currently we're in beta. Our goal is to graduate and to release the API on 1.0. We've got gotten feedback.

B

I got a lot of questions from companies with the API is in beta. Can we can we can we run our business on beta yeah? You could, but would you take the rest, the risk that we do some breaking change and some time we're making some silent some subtle breaking change?

B

So that's one risk when, when good factor on the on releasing this is that we have some internal workloads at Microsoft, some product within Microsoft that are going to use the soceity yet, which is far as really a good thing, to kind of have a good engagement with the team that we are interacting with so so. This is also one good reason for us to move to be a beta, the API on 1.0 to a stable version. So the second piece that we are bringing is we are going to bring more entities.

B

So what is coming and I should have listed it here. What is coming is we are bringing list items we are bringing lists, we are bringing drives we're bringing sites. So there are four more entities that are very sharp on Central, but these entities are coming, so you will be able to say, okay, give me Oh besides that satisfies query. Whatever including star technically- and you could also be saying, give me list items and sites that satisfies a particular query.

B

So in the SharePoint workload you will be able to specify multiple type of content and the search result will be an aggregation, a single list, a list item site whatever that you've been asking for the public ones here in terms of search capabilities. So we are adding some more such capabilities, so we are finally supporting select properties for all types. So, let's take an egg egg examples, you, you need to manipulate Drive items, but you only need two properties.

B

So by default, when you request a drive item, we we give you a list of is seven or eight properties. So if you want to trim down the list of property to optimize the content that is flowing over the wire, you now can do that you can restrict to specific properties. The second part is for type that are extensible like list item now you will be able, in list item, to specify I, want a list.

B

Item and I want custom, :, a custom column, B you'll be able to specify the custom column that your list item has.

B

So these are for select properties. Sorting is something that we didn't have previously sorting was by default by relevance, except it was by date for email and message. So now we will support the ability to sort result on any suitable property. So this is a something new in DT and I will show you how it's going to look like in the API.

B

The third source capability is refiners, so this has been a huge feedback that we've received, and that was one of the key missing feature, especially for people who have been working with sharpen search for a long time. Sharpen search has incredible capabilities with refiners.

B

We renamed them as a great, which is more seems like to be a more an industry-standard, so we have modeled aggregation in the API and I will show you a little bit later how this is going to look like in the payload, in the request, how you can express the aggregation in the request and how the aggregation will come back in the response and the last feature we are adding, which is kind of a icing on the on the cake is a did.

B

You mean spell check, feature where we will be in the back end, calling a speller to potentially to suggest some spell check correction. So today it will only work in a suggest mode will actually suggest something. It will not alter the query. Altering the query is something that we're we are thinking about. The altering the query means it will actually change the query on your own behalf and tell you by the way we've altered the query, because we know we know this. This word didn't make sense, so we change it for that particular one okay.

B

So this is what is coming. What I'm proposing to show you in the next two slide is how such will look like in the graph request. So, first of all those of you that were not familiar with the Microsoft Search API in graph, so we are having a post payload. We are a little bit different from other payloads. The reason is such easily bit specific search, you're, not specifically accessing one one item, and also we want to express some more complex things. So softly soft sort is relatively simple with. When you go to aggregation.

B

You will understand why this was not modeled as as a pure rest api as other pure rest code, where everything can be passed into the URL. So here, on the request side, you can see that you can specify in the sort flows you can sort on multiple properties, so here I'm, sorting by author ascending. So this is an alphabetical ascending and so, by date, creating they created. So I have here I'm a bib ability to specify multiple, which is where I want to solve this.

B

The underlying assumption here is that the property is suitable, so either in the connectors world or in the SharePoint world. The property has to be suitable. So this is for sort. The next one is a little bit more hairy, so I try to simplify. It is a ch3 Finers, so the cheese and one very good example, one very good in iteration is when you're requesting for file.

B

Ok, I want I, want a file that satisfies foobar, but I also want to, on top of the search result, to get the list of what will be available file types satisfying this query. So here what I'm saying is I am saying that I want the aggregation based on the property file type I want five aggregation. I want five different buckets, we'll see what those buckets are and I want to get the most represented bucket. So that's the bucket sort accounts and there should be at least a minimum of ten my items target.

B

So this is how I'm expressing the the aggregation request and then in the response, we will get the search results so that there is no change. So it's important to note that for aggregation, you get the search result, but on top of the surface you get something else you get another table and this other table will list you the list of buckets.

B

So here it's saying that they are actually 289 documents, satisfying this foobar query and that has the file type equals talk x and then 45 documents satisfy the query and have file type, equal, doc and so on, and this numbering is actually sorted by from greater to to lowest in the buckets you get a token. So for those who are familiar with SharePoint, that's the way SharePoint works.

B

So this token is actually an uncoded string so that, if you actually want to filter the results now, I have all these documents and I see that there is 289 Doc's, 45 dogs and maybe more and I want to drill the and just look at the docx one. What that which you would have to do is in the saw in them in the query, clothes and is passed the refinement token alongside and the query string, and this will trim down the results to only the documents that are that have file type equal to X.

B

So this so aggregation has been something that is, that is I've been asked a lot. If some people have some feedback feel free to reach out to me offline or in the chat window on this particular one. We've been ruling and iterating a little bit on that in shop. What you could do really a lot of things, so we are trying to rationalize a little bit come to what the industry is seems to be doing.

B

We've been looking at our other API the surface, to really see what what would do really people need and we've been trimming most of because I'm in shape on there's some crazy things you could do is refinance. We are kind of shaving some of this feature away. That's.

A

B

A

It's really good to see this much improvement coming through as well on this IPR, because I know in the past, especially their existing ecosystem. It's been something that's been very, like heavily used, I quit question. He was like what kind of scenarios do you see this being like useful within existing applications, so.

B

This one I guess, if you want to build a decent search experience I mean the ability to refine on file types. If you want, if you're building a so the scenario was talking about, is if your um your company has a custom app for employees, I think I can imagine and I know that large companies I build their own, my Microsoft whatever, and so, if you want to expose search for this in these in such an experience, you need a refiners. That's the basic thing that customers won't want you.

B

When you are looking for file, you want to be able to refine by typically author file type and probably size. So these are. This is one one example. I know in the SharePoint world, people are using refiner for some crazy things. People have even building a ecommerce site based on that capability.

A

Yeah and in the past, the implementations I've had and we have best from the cause. Well, who's, probably kind of roll out a bunch of them is you know in a company where there's lots and lots of site collections being I, just use, search to find the correct site collection. Now the indexes are so up today or even span list items across multiple lists. It's a good way of just kind of getting all that information in one hit, rather than having to query the list API for every list within every site, collection that comes through.

A

So that's really neat and, and one other thing was what we used it for site collection fetching. Now there is a site's API in the Microsoft graph as well. Now that is up only because it's not security, trimmed, and so there is a scenario that was used for search that it can now be done with the site's API as well.

A

So as you're looking at things is really worth reading the top-level pages in the learn section of our Docs, because it explains the most common scenarios and why you'd use them to kind of introduce you to these API. So there's just a bit of a tip there, because the graph is growing very very quickly as people like Nicholas are rolling. These features out. So that's great yeah.

B

And by the way, these everything I've shown here is not up only its delegation permissions so yeah this everything gets executed on the behalf of the user and that's what we're prioritizing for. We really want to build such for end-users and.

A

So, in terms of what goes v1 will that be everything that's existing Leon, the beta API will go v1 and these new teachers you're talking about war being beta, I, guess.

B

What I expect is that refiners and sort my still be in beta all the rest will be in v1. That's that's my expectation, yeah and the reason is we want to still have a little bit of room to potentially make some adjustments, some API changes in the API ships. So that's what we want to, but this is in beta, but all the rest will be available in in release.

A

See any other questions in there, so Nicholas thanks so much for today. That's awesome. Thank you for jumping on and if there are a lot of questions you say we can collaborate on those in the chat window after the fact as well.

A

Thank again for jumping on this morning, mate and hope everything is what we okay, so next up, we have only show, is actually in our team on the developer experience side and has been working on the changelog, which has been something that has we've had a lot of feedback on in the last few years of how do we keep up the date with the graph so Molly? Thank you for joining and if you want to take over the presentation, so you can share your screen and show it in action.

A

I'm really looking forward to seeing all the feedback on this.

C

Ok, I did not get the memo about the aesthetic for the presentation, so you will all be treated to my favorite animal stock photos available and also to see the wonderful changes that we've made to the change log. This is sort of our before view. This is the status quo, the current change log that we have, and then this is after a redesign, as you can see it's cleaned up.

C

We have a few different use features on the page which I am will be more than happy to walk through okay, so things that we've added our filters, search and sort an RSS feed and, on the back end, we've done some automation in order to decrease the amount of errors that we see when you are looking through your changes, so we have a variety of different kinds of filters. We have product filters for like what API you're using or you're using SharePoint. Are you using mail?

C

So that way, you can see changes to only the api's that you happen to care about similarly version filters, so whether you want 1.0 or beta, you can now just only see that you can check to see what's available on your stop and cloud choice, and you can also look for specific kinds of changes. So, if you're only looking for subtractions or additions, we got you covered. We also have an RSS feed in case. You need more notifications in your life. I know.

C

I do but also it's helpful, because you can see changes that are just by whichever category you would want to filter off of. So, if you just want to see what is relevant to your sovereign cloud or you want to see what is relevant to you know a specific API, then you have the ability to do that now. I felt like search and sort were pretty self-explanatory, like you can search through it, you can sort through it.

C

So I could make a slide, but I did make one for automation, which you do not necessarily need to know the backend details of to know that it sort of will end up solving a variety of problems.

C

Specifically, if somebody is taking their somebody moves from beta to 1.0, they oftentimes will just copy paste that bit of the they will copy paste, the change that they made for beta, and then they will paste that into the 1.0 documentation, but it will link to the wrong documentation and well oftentimes, the just, not correct, and so in order to decrease that, and also to make sure that we are showing all changes that occur. We have now automated the way that it works. So that way, you will make sure that all the changes are complete.

C

All of them will link to the correct and relevant documentation, and it should be a lot more smooth. Does anyone have any questions, and or would you like to see a demo sure.

A

On the demo, male you're teasing them all come on, I mean the Penguins cute for sure.

C

This is it, this is all sample data, so it is, there's not that much there and you'll see that, like a lot of these have no things, but you just want users changes. You just want additions. So far, all the sample data that we have is global service in 1.0, but it's quite exciting. We have a lot of different categories and sub-categories that you can look under so that way you can get precisely what you want over here. We have the RSS feed. Here's our sorting.

C

And then you can hide or should filters and let's search for something, let's search for photo.

C

There are any questions, a.

A

Lot of people saying this is great in the chat in terms of time, saver and happy. We kept up to date, which is great I will say we did get a lot of feedback on this and it was great to get the features and the ways that people would want to see this data. So that's great, and so thanks very much for driving this model. This is excellent. There are, there are future things that we will be doing around the Developer Program with this too.

A

So this is a bit longer down the line, but now we have this in a structured way before just so, you know it was marked down our PM's when they would push their doctor. Angel ooh like Nikolas and Phoebe, would have to go and manually write these change log entries, and so we had some formatting, but it wasn't enough, and so now this has been moved back very, very close to the service when it's deployed and also generated, and the authors have the ability to make those things changed.

A

But what we want to do with the Developer Program is if we can see what API you're using will actually personalize this change log for you as well. So this is just the first crack at this, but, as you can see in Molly's address bar, you know you can bookmark these things and keep coming back based on the queries that are happening or you know, go grab the RSS feed for that particular search result and be notified when new things pop up.

A

So when will this be released? Is the big question that's coming over on the chat Molly? It.

C

Is not released by the end of the month. I will be personally very sad. It's really close, it is just Gouda chrisley, close yeah.

A

This has been a very big project, with lots of lots of people involved so, but it would be great to come out. You will notice that the link will change that euro is actually internal test link, but it will be on a developer top Microsoft comm site now and not in the docs, but we will handle a bunch of redirects so that, if you go to the existing change, log you'll immediately be punted straight over into here too, and, as you can see, it will work in both dark and light mode.

A

Now, one thing to note: in the bottom right there you'll see a would love your feedback, which is actually a way of collecting in SAP or developer. Satisfaction from you, our developer community, and so it encourage you when this new system is live to.

A

Let us know what you think by filling in those forms when they pop up, we do actively sit down and look through all those ratings and even the ad hoc feedback, and if you leave your email address there with gdpr reasons, we have like 30 days basically to reach back out to you and get follow-up questions if we have them based on the feedback you've given. So we are actively listening and I think this is great evidence of what Molly's been driving here. Is the PM on this and to get us done so.

C

A

C

Think I need to give you back control now. Yeah.

A

Thank you money. That's awesome. Don't.

C

A

So I'm going to share my screen again and then for the third topic. Today we have Phoebe who's, our p.m. for the core one of the PM's from the Excel API site inside of Microsoft graph, and wanting to share some updates on. What's going on on the Excel API is on Microsoft graph, so Phoebe do you wanna take over and present your part of the session thanks for joining us? Well, because I know time, zone wise. This is a bit rough on you, based on where you're located.

D

So hello, everyone, I'm viviane from the excel graph theme and the Excel graph team locates in Beijing, and now it's almost midnight but good morning, everyone and our team major working on Excel, REST, API and I'm very happy to have the opportunity to share some update from excel graph. In today's call.

D

So, as you know, Microsoft a graph provides a unified programmability model that you can use to access the tremendous amount of data in Microsoft and our excel graph is one of them and you can use Excel REST API s to extend the value of your data to do analysis, performing calculations generating report storing or tracking data. Actually, we have a lot of customers who are using excel graph as the solution in their tailing work.

D

So you can see from this slide below that the following companies or teams are our existing customers so for more details about all available excel graph api is, please, go to our graph documentation site to see how to easily play with them.

D

So for what's new, we have two items to share today. The first is long-running operation pattern and a async create obsession API, which uses this pattern. We decided to invest in this area because we noticed there were telemetry data showing some failures caused by we passed the time out, and we also heard from our customers reporting failures. For the same reason, so we decided to introduce the long run operation model to address this issue, in that we could provide a more reliable, Excel graph service to our customers.

D

We also complete one API in this model, who are created a possession creation scenario which I will introduce with more details in the later slides and the second update will be. We will provide more detailed information in excel arrows. We receive the feedback that arrows are confusing, sometimes in self and Excel graph didn't provide enough guidance on how to handle arrows so the customer feedback. We always paid a lot of attention to so to address the usability 10 points. We will make improvements and provide more detailed error information coming future.

D

We will have. We also have documentation plane soon to introduce the error handling part so for the first update item long-running operation as Excel rest, it records need to be completed within certain amount of time. Otherwise, the request, the world time outs and users will see the reflector failure. However, we noticed some api's actually reply, indeterminate I'm to complete.

D

So, to solve this easier, we introduce the long run operation pattern and by adoption this pattern, our users do not need to wait until the action is completed before returning a response or worrying about the timeout for the class, and currently we have only one API using the this patent. The long run operation pattern- that is the asean-korea session API, which is already in beta version, is month. This is a good news right and in future we hope we could provide a more and more api's with the long run, operation pattern.

D

Ok, so next slide I will give some detailed introduction to the async read create session API. The first step is to initiate initial, create a session request, so in a request at a header of the prefer response async to indicate this is a long-running operation request and just start assuring in the below. There might be two possible responses. The left side is the common response. Will two or one created status code, so this will happen when the workbook could be open a short time. It won't enter a long-running operation pattern.

D

Instead, it will return result directly very shortly. The right side actually is the example of response for a long-running case response will return the to 202 accepted status code. So if you see the 202, it is for sure you are entering a long-running operation pattern with async needed information provided afterwards, and the next slide is the second step to pull create sessions Taylors through getting the operation ID method. There are three possible seriously here.

D

You can see the green highlighted in this light, that that are the brand-name subsidies and the fail status so for running it means the creation is still ongoing and the succeeded means the session creation is successful and the failed of course means the creation doesn't succeed. So, in the succeeded case you can see the resource location will contain a detail. The URL that created session information- and one thing needed to be noticed in the response here is in the failed case, because the async request is successful.

D

It will return the 200 ok in the HTTP status code. However, the creation operation is spelled so the status will show failed in the response. It is a little bit different to compare him to the positive responses in the excel graph api's and in the last step for the ACN create session. Api is to acquire the session info. You can use to get a request to get a corresponding session information to gather information and in response you could get a created session ID through the ID value here and now.

D

So the previous three slides show them to end. The basin creation process is completed.

D

Ok next update I want to introduce the improvement we made to provide more detailed information for Excel errors, and you could see the comparison between before and after for every object in cell graph and in the past, error object has properties of error, code, error message and in error and in a coming future we will make some improvements and provide a 3 layer structure in our error object, and you can see the example here that there will be top level second level and the third level error errors in the error object and for the detailed documentation about arrows.

D

We will be published soon and please stay tuned and check on updates. In the dockside and to save time in today's cowan to be simple in this slide, here is the brief guidance on how to handle excel graph arrows. The most straightforward step is the client should expect it to act according to the instructions provided in the second level error code, and the message is first on.

D

Ideally, we will provide enough information in this level, but if there is no effective instructions available in the second level in our published Docs, so this chart on the top level code and the messages, if, unfortunately, they are still knowing affecting destruction available for the top level, then you can check on the HTTP status code to know the information about error and please notice that the level code and the messages developer use only example for debugging purpose.

D

That's the two updates I have for today's from excel graph site. We also wanted to hear more feedback from our customers from you. So if you encounter any issues while using the Excel REST API eyes, please raise questions in Stack Overflow and we provide a link in this slide and please please remember, to use the Microsoft a graph excel as the tag. So we could response in time and if you want to have any specific feature that is not available in Excel raptly also, please also send a request to the user voice Channel.

D

We also provide a link in the slide and Thank You Jeremy to providing the link in the chat chat channel and at last please also check on our documentation in future for more updates about excel block. Okay, that's my part for the update today.

A

D

Question yeah Thank.

A

You Phoebe and again pretty sure you being on satellite for this, but we get a lot of demand for this topic, so it's great to have lots of questions different in different forums. I mean they're, very active mistake for sure, sighs put that link in there. Martinus had a question. We said whenever an excel session non-persistent one and calculated it with certain values, I need to download the end results at the PDF file. Are there any plans for this?

A

The only way I can do this now is to create a persistent session then use the files API to convert to PDF. However, this is impossible for many years of scenarios, always in the same excel file since I want the original file to remain unchanged.

D

Yeah, actually, we have non-persistent model for excel api's mm-hmm.

E

D

I'm not sure if this can save the efforts for this customer yeah.

A

Maybe this one is something we could take offline potentially into a flow question, and then you can help help there with the customer yeah yeah. That.

D

Overblow question to us and our future crew can provide a more detailed, a solution. Yeah yeah.

A

Yeah that one seems quite complicated and what is the difference between a non-persistent and a persistent excel session? For those that don't know, phoebe.

D

Soulful persistent a session, all the changes made it who that excel api's will be saved it to the document. The non.

E

D

Api's, you can do some of the calculations as a calculation, but the changes won't be saved it to the file. That's the difference between these two modes right.

A

Yeah, so in the case, where he's trying to create a PDF, if it's not persisted back, the PDF won't reflect those changes that have been made.

D

But the PDF will fire I'm, not quite sure, because most of the api's we are working on is communicating with excel files with mr. Todd's Excel Oster Asda file formats and double-check offline. If there's any potential solutions there yeah that.

A

One in stock for sure yeah.

D

Thank you for the question. I really.

A

Appreciate your time and it's great to see more coming through on the lawn operations, I know there's many api's on the graph that could probably do with that. That I know we've had lots of feedback from various people on, and so our API review board at the moment is trying to make sure consistently we're using the same schema for that long operations, and so you're gonna see that pop up in various places as well.

A

So- and this is just another example of where we're doing that because of the notion of what those api's are doing under the covers and the fact that it can't just easily in a lot of cases, you know respond with a across a request. So thank you again Phoebe.

A

So, just in terms of other news, we have had at a program a technology adoption program for Microsoft graph for quite a few years now, but we haven't been able to scale it because we didn't have the resources to manage something of the scale we'd like to have done, and in in recent weeks we are growing our our customer partner experience team, which I'm gonna actually be leading I'm going to be handing off the developer experience work, so the graph Explorer changelog Docs SDKs will actually be headed to a new lead within yoona yoona terenas team, and I'm going to be driving up that directly and brian who who's been on.

A

So many things already around graph and what kind of headed up the 30 days of graph blogpost in 2018 and has never really been out an escape. The the cause of the graph is our first member of that CPX team, which is great now. The reason I wanted to bring this up today was is we are growing that fellowship at the graph, our technology adoption program, it's been primarily MVPs and regional directors and and select existing partners in the Mike's.

A

The 365 ecosystem have worked with us over the years with SharePoint and exchange, and so now we're opening that up to more partners and sis that are already under NDA. And so, if you are interested in joining our tap program, we do, though we will be doing those calls monthly on the first Wednesday of the month at 8 a.m. so. In most cases it will be the day after the community call, but this month being the first one we did kinda deadly it wasn't.

A

It was last week because of the way that the month rolled, and so in those monthly calls. We do a similar format to this, but we just do it on an earlier crank. So you know, Nicholas would have presented some of this stuff, maybe six months ago to the tap scene, where he would have explaining what they were considering, building and the design and so forth.

A

And so this is a great way for partners to see what's coming, if they're doing kind of strategic playing in their own products to know what's going to be out there and these community calls are extremely useful for people to understand what you know what's just being released or what's just around the corner and what's public and so that the tap programs are definitely a little bit more behind the curtain.

A

A little from that aspect, and it will give the opportunity for people in the tap to have access to private previews of API so on in beta API endpoint. Yet, but you know, will we announce those things and partners are able to kind of sign up for those as well? And then the goal of this really is to make sure that we work closely with our partners out there to make sure they're successful in app source, and we are looking for.

A

You know really good case studies and keynote showcases, so every year at build and ignite we're, always looking for new things for Rajesh yaar and such and Adella to present on stage regarding Microsoft graph, and so again, if you fill out that form, we will review their client IDs just to look at current usage of your applications. There are some criteria from a bar perspective that we're setting on that, and hopefully we'll see you in there as we do on this community call so into addition to that.

A

As I mentioned the beginning of this call, there are a variety of other calls that exist there. Sharepoint calls there's a normal SharePoint developer, call a general developer one and a SharePoint framework, one that better even runs its on the call, as well as a variety of own ones, across kind of Microsoft teams, office, add-ins, power, apps and the identity. Nap adaptive card so definitely check those things out and you can add, the ICS is for those into your calendars.

A

Whatever culinary system you use, hopefully you're using Outlook and Exchange, and so that's the easiest way to kind of keep up to date with that and then there's a podcast that currently Paul I doing. We had some really good episodes recently with variety of different internal PM's.

A

If you haven't checked out the fluid framework, podcast yet definitely go check that out fluid is going to be extremely hot property once that is open-source publicly and saying does a really good job of explaining what that is, and as soon as it's available publicly we'll be sharing some of that stuff on the community call here, as well with the graph context.

A

So with that all of these videos are being recorded right now and they're published on the Micra 3-6-5 developer, youtube channel and I highly encourage anyone. That's on Twitter at Microsoft, 365 dev is the Twitter handle that we communicate for everything through Mike's than 365 dev, sharing news blogs, all the YouTube recordings, any upcoming events and so forth. So definitely encourage you to do that and then our next call it's always the first Tuesday of the month and so August 4th 8:00 a.m.

A

PST and again that URL is the one that will get you the ICS and your calendars. So you don't forget and I'm always kind of posting on Twitter to remind people the day and the morning of to come, join the call so that I'll just see if there's any questions on here any hands up with regards to anything else, because we do have two minutes: do they even know we got through three speakers. So thank you for the speakers for being on point today. Okay, well, I was joking rid of it, so Jeremy.

E

From Gary any news on the to-do API not much activity since build yes,.

A

So I actually chatted to them you yesterday they are on track. They said August September in the slide that they announced at build. I'd originally thought it was earlier than that quite frankly, but August September is a timing, so next month, they're looking to have that in preview in the beta endpoint for to do so. Unfortunately, there's nothing quite out there, yet we we're working with them to get them released into the service. In that timeframe it looks like that's definitely gonna be here. They are pretty much ready, there's just a few things.

A

They're setting up, but them yeah come come. August I would expect to see them on this community core next month. Gary another.

E

Question from Jo Paula, some of them come in. Will there be any openings for the dress costume where.

A

E

A

Nicole is all on here today and I would scale the Oscar he's actually wearing is Geoff's it right now.

E

Unfortunately, not right now sorry, how's.

A

The lockdown hair going Nicola, hey.

E

If it keeps growing, I'll see how I can go.

A

I gave up the ghost on that and just shaved it all yeah.

E

You're, a quitter I can tell.

A

And then what other questions are there more elaboration of the recent throttling announcements? Yes, so one thing of note is: if you look at the FRA tling page, which is in the graph Docs, which I would just share on my screen,.

A

Vincent Bray, who I don't know if he's on the core it's hard to scan through all those names, but in the docs. If you look at the docs here and you go to use the API on the develop section, you'll see we have this Watling page and for a long time, we've had feedback that it's very hard to plan design, an application. If we don't know how to best, handle the api's.

A

And so you will now see that, if I scroll down to service specific limits that we actually and not just got the outlook API limits in here, which has been there for probably over two years now and they've, been here since I joined the team. But we specifically started reaching out to the different workload teams and validating what the particular limits are for each of the workloads.

A

Now you know, ideally, we would want the same throttling limits across all of the graph, but in reality, because we are essentially a federated api across 50 plus workloads across Microsoft. The underlying services often have different different behaviors because of the notions of how they're being used, and so Vinson has been going through and essentially listing.

A

All these for all the different API is to explain the exact limits, including the Excel API, is that Phoebe presented on today, and so, if you haven't had a look at that, I would absolutely go and check out that new list, because it does often help you know we do state.

A

You know that the SDK will handle those 40 nights for you and back off when those calls there are often times where you're just trying to do planning and estimation of like how long it would take to do certain items and understanding where those limits are in terms of how many requests you can make it is, is really really important and.

E

One thing to mention on the throttling is that this is very much a work in progress, updating all the various workloads on there and yes, they're aware that we have now exploded that the amount of information there they're gonna be working on the navigation to help make this easier to browse through. So it's not just you know, implement scroll going down the page on here. Yeah.

A

Jay in terms of the Microsoft engineer, in brackets being corporate management. Quite frankly, what we've tried to do is not put product names in here at all, so you know: even teams is team work because it's more of the productivity angle that we're discussing and Excel is workbooks and charts. So that was just one oversight where we actually had a product name in here in the graph and that's really the way we kind of approach.

A

This is that when I talk about work, clothes or Brian or anyone else does, these are the kind of the workloads that we're talking about the left-hand side of this learn. So when you're, exploring things like cell starting on the overview page here is gonna, explain to you like the basic high-level of what this this workload is meant to be about on the graph and same with things like Nicholas's search. Is this will explain?

A

You know why why you would use search within Microsoft graph, so these topics are useful when again, they're not meant to be product orientated, so yeah I'll leave it at that in terms of reading the tea leaves on and on those things, excellent. So there any other questions if limits are appurtenant. Does that mean that a tenant with 20,000 users has the same limit to a tenant with 10 only 10 unit users?

A

Yet that does so that they're very literal in the way that the the FATA lling limits are, and so you know, regardless of the tenant, as you mentioned, being 20,000 users or 10 users, those limits will have to be taken into account. The majority of the Outlook Exchange Online ones are basically around the individual mailbox limit. So in their instance, if there was 20,000 users in a ten or ten, the it's a it's done at the mail, but the user level, so it wouldn't be impacted by the size of the tenant.

A

But there are certainly some on here that have that limits of more at the talent level throttling than it is based on. You know how many times you're aggressively hitting an individual's mailbox to maybe pull down. They calendar information. So there's just something to be in mind of as you're using those api's.

A

You know use the SDKs watch the four two nines and then you can actually look at this detail to get a little bit more granular on individual work, clothes and what those kind of behaviors are so yeah in the use your API section and under throttling is where all that information is obtained.

A

Okay right time, so thank you again to Molly Beebe and to Nikolas for presenting today the recording will be posted in the graph blog. You want to check that out, definitely plug into the process feed for that. So you get those notifications as well as all the changelog notifications when we go live with that, and thanks again thanks Brian for chiming on with reading the questions out and putting information in there too. So have a good week. Everybody and see you next month.