If anything’s over, it’s the city previously known as “Instagram New York.” Covid-19 has pulled the filter off much of o

Why is pyodbc trying to convert something from varchar to numeric?! When I commented out line 14 in order to use cursor.executemany() without the fast_executemany feature, the script worked just fine! I was able to insert my data without any issues.

http://live-stream.munich.es/twr/videos-Celtics-Heat-v-en-us-1tuh-.php

http://live-stream.munich.es/twr/videos-Celtics-Heat-v-en-us-1sbc-3.php

https://assifonte.org/media/mln/video-Cleveland-Cavaliers-Magic-v-en-us-1kug-16.php

http://news24.gruposio.es/ktn/Video-Rockets-Pacers-v-en-us-1zxs-12.php

http://news24.gruposio.es/ktn/videos-Cleveland-Cavaliers-Magic-v-en-us-1dzq-.php

http://live-stream.munich.es/twr/Video-Celtics-Heat-v-en-us-1pxd-18.php

https://assifonte.org/media/mln/v-ideos-Cleveland-Cavaliers-Magic-v-en-us-1mqy-7.php

http://go.acaps.cat/vac/video-bragantino-v-sao-paulo-v-pt-br-1kkq2-3.php

http://news24.gruposio.es/ktn/Video-Cleveland-Cavaliers-Magic-v-en-us-1qto30122020-2.php

https://assifonte.org/media/mln/Video-Wizards-76ers-v-en-us-1gzr-.php

http://live-stream.munich.es/twr/Video-Jazz-New-York-Knicks-v-en-us-1xir-.php

http://go.acaps.cat/vac/Video-bragantino-v-sao-paulo-v-pt-br-1khr2-6.php

http://news24.gruposio.es/ktn/v-ideos-Cleveland-Cavaliers-Magic-v-en-us-1knx-24.php

http://live-stream.munich.es/twr/Video-Jazz-New-York-Knicks-v-en-us-1qtk-10.php

https://assifonte.org/media/mln/videos-Wizards-76ers-v-en-us-1tmu-7.php

http://go.acaps.cat/vac/video-Bragantino-Sao-Paulo-v-en-gb-1uwv-.php

http://news24.gruposio.es/ktn/v-ideos-Wizards-76ers-v-en-us-1vvf-.php

http://go.acaps.cat/vac/videos-Bragantino-Sao-Paulo-v-en-gb-1coo30122020-13.php

http://live-stream.munich.es/twr/videos-Jazz-New-York-Knicks-v-en-us-1dco30122020-23.php

https://assifonte.org/media/mln/video-Wizards-76ers-v-en-us-1twg30122020-12.php

one of the most insidious types of concurrency bugs because they may not show up until years after the code has been placed into production. They are usually precipitated by a change in the environment the code is executing in, or an unprecedented occurrence. In these cases, the code seems to be behaving correctly, but in reality, there’s just a very high chance that the operations will be executed in order. Sooner or later, the program will have an unintended consequence.

In order to load this data to the SQL Server database fast, I converted the Pandas dataframe to a list of lists by using df.values.tolist(). To ingest my data into the database instance, I created:

3. Long to-do lists are fine as long as you set deadlines. Lots of people have different opinions on this one. Some believe if something has been on your list for more than X days, you should either do it or delete it. Others suggest your list should have a fixed amount of items, and you have to either delete or complete something on it before adding any new task. I also know people who use only daily lists and won’t even track anything they’re not focused on accomplishing that day.

In summary, I was able to fix the “Error converting data type varchar to numeric” by converting my float column to string with exactly the same decimal point number as defined in the SQL Server table. It was quite surprising to me that pyodbc doesn’t handle that under the hood (or maybe is it fixed in more recent Pyodbc versions?).

After fixing the issue, the script ran 100 times faster as compared to running it without line 14 (cursor.fast_executemany = True). Note that it’s so fast because it loads the entire data into memory before loading it to SQL Server, so take loading in chunks into consideration, if you come across out of memory errors.

Prioritize what requires your input for others to move forward. Think of it as “downstream” impact. Is your lack of action on a task blocking others from moving forward on something that is important to them? Move it to the top of your list. Especially if my involvement doesn’t require much time, I try to move these tasks along quickly so as to not create process delays. If I know it’s not something I can attend to quickly, I’ll give the dependent parties a clear timeframe so they can gauge their own plan or come up with a different solution. And if I don’t think I need to be involved at all, I’ll extract myself from the project completely.

After managing large teams across Google for many years, and now, as an investor, having the chance to observe dozens of highly effective startups, I’ve seen there is one major difference between people who are “just” productive and those who are truly effective: knowing how to prioritize.

My own relationship with prioritization is a perpetual work-in-progress. I still fall back into a habit I’ve long tried to break: working on the stuff that seems easiest to complete, or the tasks that give me the most satisfaction or accolades, or simply the latest item that came in, rather than what I actually most need to work on. But I have been improving lately by following three guidelines, which might help you, too.

I’ve been recently trying to load large datasets to a SQL Server database with Python. Usually, to speed up the inserts with pyodbc, I tend to use the feature cursor.fast_executemany = True which significantly speeds up the inserts. However, today I experienced a weird bug and started digging deeper into how fast_executemany really works.

My dataframe schema: summertime bool time datetime64[ns] unique_id object measurement float64 entered datetime64[ns] updated datetime64[ns]

In summary, I was able to fix the “Error converting data type varchar to numeric” by converting my float column to string with exactly the same decimal point number as defined in the SQL Server table. It was quite surprising to me that pyodbc doesn’t handle that under the hood (or maybe is it fixed in more recent Pyodbc versions?).

It’s really no surprise, given all this, that the last viral “New York is dead” cri de coeur came in the form of an opinion piece in the New York Post by a wealthy white man (no, not the president). New York may have been declared dead in the past, he conceded, but this time, it’s “completely dead.” Or that a letter sent to the mayor by 163 business leaders in September that described “widespread anxiety over public safety, cleanliness and other quality of life issues that are contributing to deteriorating conditions” was barely recognizable to those actually living here; Kathryn Wylde, the president of the Partnership for New York City, told the Times they had waited to send it till after Labor Day because “they felt it was unseemly to be writing from the Hamptons.”

2. Think “top down, not bottom up.” Whether you use your inbox, a to-do list, Post-its, or your memory to track open items, there’s one mantra I want you to repeat: “Top down, not bottom up.” Do what matters first in terms of impact. I use the metaphor “avoid snacks,” which you can read more about here.

This error was extremely confusing to me since the data types of my Pandas dataframe matched perfectly with those defined in the SQL Server table. In particular, the data that I was trying to load was a time series with a timestamp and measurement columns some metadata columns.