Can someone please help me format the visitStartTime field. How would you transform 1499484877 into a time?
", "messages": [ { "id": 389264, "postDate": "2018-09-18T12:57:02.930Z", "content": "This is how to do in python. I'm not sure how in R.
\n\nimport datetime as dt
\n\ndf['time'] = df.visitStartTime.apply(lambda x: dt.datetime.utcfromtimestamp(float(x)))
", "rawMarkdown": "This is how to do in python. I'm not sure how in R.\n\nimport datetime as dt\n\ndf['time'] = df.visitStartTime.apply(lambda x: dt.datetime.utcfromtimestamp(float(x)))", "votes": 7 }, { "id": 389253, "postDate": "2018-09-18T12:26:02.850Z", "content": "Can someone please help me format the visitStartTime field. How would you transform 1499484877 into a time?
", "rawMarkdown": "Can someone please help me format the visitStartTime field. How would you transform 1499484877 into a time?", "votes": 4 }, { "id": 391279, "postDate": "2018-09-21T14:16:40.807Z", "content": "I am simply using as.POSIXct(visitStartTime, origin=\"1970-01-01\", tz=\"UTC\"), and once you get the hour, minute and seconds extracted, I also find it useful to have a \"continuous\" time feature each day by doing something like (hour + minute/60 + seconds/3600).
", "rawMarkdown": "I am simply using as.POSIXct(visitStartTime, origin=\"1970-01-01\", tz=\"UTC\"), and once you get the hour, minute and seconds extracted, I also find it useful to have a \"continuous\" time feature each day by doing something like (hour + minute/60 + seconds/3600).", "votes": 2, "replies": [ { "id": 391307, "postDate": "2018-09-21T15:20:18Z", "content": "Hey yilisg,
\n\nI'm not too sure what you mean but I take that you compute some ratio that tells you how far in the way we are at? I did the same thing with the following piece of code.
\n\npython\ntime = df['visitStartTime']\nprogress = (time.dt.hour * (60 * 60) + time.dt.minutes * 60 + time.dt.seconds) / (24 * 60 * 60)\n
This returns a value between 0 and 1.
", "rawMarkdown": "Hey yilisg,\n\nI'm not too sure what you mean but I take that you compute some ratio that tells you how far in the way we are at? I did the same thing with the following piece of code.\n\n```python\ntime = df['visitStartTime']\nprogress = (time.dt.hour * (60 * 60) + time.dt.minutes * 60 + time.dt.seconds) / (24 * 60 * 60)\n```\n\nThis returns a value between 0 and 1." }, { "id": 391314, "postDate": "2018-09-21T15:30:44.593Z", "content": "I guess that works too. For me it's just a number between 0 and 24. Sry about the typo above.
", "rawMarkdown": "I guess that works too. For me it's just a number between 0 and 24. Sry about the typo above." }, { "id": 391352, "postDate": "2018-09-21T17:06:20.743Z", "content": "No worries! I just wanted to make things clear :)
", "rawMarkdown": "No worries! I just wanted to make things clear :)" } ] }, { "id": 391099, "postDate": "2018-09-21T09:11:38.317Z", "content": "Even quicker: df['visitStartTime'] = pd.to_datetime(df['visitStartTime'], unit='s')
There is a problem with pd.to_datetime. It does not interpret the timestamp correctly.\nFor this reason its better to use the external datetime module.
Can you be more specific?
", "rawMarkdown": "Can you be more specific?" }, { "id": 391110, "postDate": "2018-09-21T09:26:52.197Z", "content": "Sorry for the confusion. You are totally right your solution is pretty neat.\nI tried the same without specifying unit=\"s\" and got the wrong dates.
Cool! You nearly got me to doubt ;)
", "rawMarkdown": "Cool! You nearly got me to doubt ;)" } ] }, { "id": 391081, "postDate": "2018-09-21T08:43:00.897Z", "content": "For R, I would use lubridate library.
The result of the code below matches with that of Epoch Converter :)
\n\nlibrary(lubridate)\nlibrary(tidyverse)\ndf %>% mutate(visitTime = as_datetime(visitStartTime))\n", "rawMarkdown": "For R, I would use `lubridate` library.\n\nThe result of the code below matches with that of [Epoch Converter][1] :)\n\n
library(lubridate)\nlibrary(tidyverse)\ndf %>% mutate(visitTime = as_datetime(visitStartTime))\n\n\n [1]: https://www.epochconverter.com/" }, { "id": 389734, "postDate": "2018-09-19T06:27:47.967Z", "content": "
Hi Joash, You can use anytime package ( Here) in R to extract Unix/Epoch timestamp to datetime format but \nR conversion does not seem to match with Micheal's command in python. I checked it with Epoch Converter, the results from this link matches with Micheal's command. So it is safer to use the python command.
", "rawMarkdown": "Hi Joash, You can use **anytime** package ( [Here][1]) in R to extract Unix/Epoch timestamp to datetime format but \nR conversion does not seem to match with Micheal's command in python. I checked it with [Epoch Converter][2], the results from this link matches with Micheal's command. So it is safer to use the python command.\n\n\n [1]: https://stackoverflow.com/questions/13456241/convert-unix-epoch-to-date-object\n [2]: https://www.epochconverter.com/" }, { "id": 390668, "postDate": "2018-09-20T15:59:02.370Z", "content": "I recommend you run this line right after.
\n\ndf.time = pd.to_datetime(df.time)\nThis way you make sure pandas converts the time column to dtype: datetime64[ns], which opens up some handy featrures.
Thank you very much!
", "rawMarkdown": "Thank you very much!" } ], "comments": [ { "id": 389264, "author_name": "Michael Maguire", "author_url": "", "post_date": "2018-09-18T12:57:02.930000", "content": "This is how to do in python. I'm not sure how in R.
\n\nimport datetime as dt
\n\ndf['time'] = df.visitStartTime.apply(lambda x: dt.datetime.utcfromtimestamp(float(x)))
", "votes": 7, "replies": [] }, { "id": 391279, "author_name": "yilisg", "author_url": "", "post_date": "2018-09-21T14:16:40.807000", "content": "I am simply using as.POSIXct(visitStartTime, origin=\"1970-01-01\", tz=\"UTC\"), and once you get the hour, minute and seconds extracted, I also find it useful to have a \"continuous\" time feature each day by doing something like (hour + minute/60 + seconds/3600).
", "votes": 2, "replies": [ { "id": 391307, "author_name": "Max Halford", "author_url": "", "post_date": "2018-09-21T15:20:18", "content": "Hey yilisg,
\n\nI'm not too sure what you mean but I take that you compute some ratio that tells you how far in the way we are at? I did the same thing with the following piece of code.
\n\npython\ntime = df['visitStartTime']\nprogress = (time.dt.hour * (60 * 60) + time.dt.minutes * 60 + time.dt.seconds) / (24 * 60 * 60)\n
This returns a value between 0 and 1.
", "votes": 0, "replies": [] }, { "id": 391314, "author_name": "yilisg", "author_url": "", "post_date": "2018-09-21T15:30:44.593000", "content": "I guess that works too. For me it's just a number between 0 and 24. Sry about the typo above.
", "votes": 0, "replies": [] }, { "id": 391352, "author_name": "Max Halford", "author_url": "", "post_date": "2018-09-21T17:06:20.743000", "content": "No worries! I just wanted to make things clear :)
", "votes": 0, "replies": [] } ] }, { "id": 391099, "author_name": "Max Halford", "author_url": "", "post_date": "2018-09-21T09:11:38.317000", "content": "Even quicker: df['visitStartTime'] = pd.to_datetime(df['visitStartTime'], unit='s')
There is a problem with pd.to_datetime. It does not interpret the timestamp correctly.\nFor this reason its better to use the external datetime module.
Can you be more specific?
", "votes": 0, "replies": [] }, { "id": 391110, "author_name": "", "author_url": "", "post_date": "2018-09-21T09:26:52.197000", "content": "Sorry for the confusion. You are totally right your solution is pretty neat.\nI tried the same without specifying unit=\"s\" and got the wrong dates.
Cool! You nearly got me to doubt ;)
", "votes": 0, "replies": [] } ] }, { "id": 391081, "author_name": "honake", "author_url": "", "post_date": "2018-09-21T08:43:00.897000", "content": "For R, I would use lubridate library.
The result of the code below matches with that of Epoch Converter :)
\n\nlibrary(lubridate)\nlibrary(tidyverse)\ndf %>% mutate(visitTime = as_datetime(visitStartTime))\n", "votes": 0, "replies": [] }, { "id": 389734, "author_name": "Prashanth Thangavel", "author_url": "", "post_date": "2018-09-19T06:27:47.967000", "content": "
Hi Joash, You can use anytime package ( Here) in R to extract Unix/Epoch timestamp to datetime format but \nR conversion does not seem to match with Micheal's command in python. I checked it with Epoch Converter, the results from this link matches with Micheal's command. So it is safer to use the python command.
", "votes": 0, "replies": [] }, { "id": 390668, "author_name": "", "author_url": "", "post_date": "2018-09-20T15:59:02.370000", "content": "I recommend you run this line right after.
\n\ndf.time = pd.to_datetime(df.time)\nThis way you make sure pandas converts the time column to dtype: datetime64[ns], which opens up some handy featrures.
Thank you very much!
", "votes": 0, "replies": [] } ], "raw_markdown_by_id": { "389264": "This is how to do in python. I'm not sure how in R.\n\nimport datetime as dt\n\ndf['time'] = df.visitStartTime.apply(lambda x: dt.datetime.utcfromtimestamp(float(x)))", "389253": "Can someone please help me format the visitStartTime field. How would you transform 1499484877 into a time?", "391279": "I am simply using as.POSIXct(visitStartTime, origin=\"1970-01-01\", tz=\"UTC\"), and once you get the hour, minute and seconds extracted, I also find it useful to have a \"continuous\" time feature each day by doing something like (hour + minute/60 + seconds/3600).", "391099": "Even quicker: `df['visitStartTime'] = pd.to_datetime(df['visitStartTime'], unit='s')`", "391081": "For R, I would use `lubridate` library.\n\nThe result of the code below matches with that of [Epoch Converter][1] :)\n\nlibrary(lubridate)\nlibrary(tidyverse)\ndf %>% mutate(visitTime = as_datetime(visitStartTime))\n\n\n [1]: https://www.epochconverter.com/", "389734": "Hi Joash, You can use **anytime** package ( [Here][1]) in R to extract Unix/Epoch timestamp to datetime format but \nR conversion does not seem to match with Micheal's command in python. I checked it with [Epoch Converter][2], the results from this link matches with Micheal's command. So it is safer to use the python command.\n\n\n [1]: https://stackoverflow.com/questions/13456241/convert-unix-epoch-to-date-object\n [2]: https://www.epochconverter.com/", "390668": "I recommend you run this line right after.\n\n df.time = pd.to_datetime(df.time)\n\nThis way you make sure pandas converts the `time` column to `dtype: datetime64[ns]`, which opens up some handy featrures.", "389403": "Thank you very much!" } }